repository stringclasses 156 values | issue title stringlengths 1 1.01k ⌀ | labels stringclasses 8 values | body stringlengths 1 270k ⌀ |
|---|---|---|---|
openzipkinzipkin | there be a time zone issue with dependent query when elasticsearch be store | Bug | the actual result be the current start and end time minus 8 hour later |
openzipkinzipkin | get trace return 500 response with response body set to responsetimeoutexception | Bug | describe the bug I m make a request to a url that look like this servicename spanname spanname endts endts lookback lookback limit limit I ve get back a 500 status code with the follow response body responsetimeoutexception it look like armeria might be time out when call elasticsearch and for some reason the response body be be set to the simple name of the armeria exception class that s just a guess though it s a really weird response body to be receive be there a way to configure the timeout that zipkin server use when call elasticsearch step to reproduce see above expect behaviour to receive a valid json response body that contain an array of trace |
openzipkinzipkin | zipkin api for each service in zipkin return total number of trace where the service be the root | Enhancement | feature a new zipkin api feature for retrieve for all service the number of trace and or span where a give serv be the root rationale when the amount of elasticsearch diskspace have grow over time e g of x gb of diskspace now remain work out which service via their traffic and sampling rate be use the most space in zipkin s elasticsearch indecie at the moment I have to get direct access to elasticsearch and run manual query to achieve this it would be great if this could be do via the zipkin api if it could be do via the zipkin api it could be do in a manual and an automation fashion for automation custom check alert and dashboard could use the datum return by the zipkin api example scenario see rationale above prior art I will reply on this issue with an example elasticsearch query for how I find out this datum currently |
openzipkinzipkin | swagger clarification | Enhancement | feature the documentation for post api v2 span should clarify that timestamp s granularity be microsecond moreover the example in zipkin lens example should use microsecond rationale currently if you add the example span and try to fetch api v2 trace it return an empty array the zipkin exporter implementation in opentelemetry cpp also make the same mistake so I assume this be an easy mistake to make example scenario prior art link to prior art more link |
openzipkinzipkin | add docker image for arm v7 32 bit | Enhancement | feature arm64 be support but not arm v7 32 bit please add support for arm 32 bit rationale there be many user that run on 32 bit os for arm even when the hardware support 64 bit example scenario people with raspian os 32 bit try to install dapr as per comment in this issue issuecomment 757541930 prior art support for arm64 be add here but 32bits be not include |
openzipkinzipkin | blank page after press the query button twice | Bug | my version be 2 23 2 when I press the query button twice then the ui show blank page below be error log typeerror can not read property reduce of undefined at discoverpagecontent tsx 360 at object ha as usememo react dom production min js 162 at t usememo react production min js 25 at discoverpagecontent tsx 357 at hn discoverpagecontent tsx 505 at xo react dom production min js 153 at da react dom production min js 175 at ms react dom production min js 263 at cu react dom production min js 246 at su react dom production min js 246 |
openzipkinzipkin | add support for remoteendpoint in transport client | Enhancement | please first look at exist issue to see if the feature have be request before if you don t find anything tell we what problem you re try to solve often a solution already exist don t send pull request to implement new feature without first get our support sometimes we leave feature out on purpose to keep the project small feature description of the feature often it be useful to add remoteendpoint to the client span specifically http request currently nor the client or the transport support this functionality the most logical option that come to my mind be that we add an option for both rationale why would this feature help other besides I example scenario what kind of use case would benefit from this feature prior art link to prior art more link |
openzipkinzipkin | flaky test in deduplicatinginsertt exceptionsinvalidate enqueue | Bug | l2575 java expect actual not to be empty at zipkin2 storage cassandra internal call deduplicatinginsertt exceptionsinvalidate enqueue deduplicatinginsertt java 90 at java base jdk internal reflect nativemethodaccessorimpl invoke0 native method at java base jdk internal reflect nativemethodaccessorimpl invoke nativemethodaccessorimpl java 62 at java base jdk internal reflect delegatingmethodaccessorimpl invoke delegatingmethodaccessorimpl java 43 at java base java lang reflect method invoke method java 566 at org junit runner model frameworkmethod 1 runreflectivecall frameworkmethod java 59 at org junit internal runner model reflectivecallable run reflectivecallable java 12 at org junit runner model frameworkmethod invokeexplosively frameworkmethod java 56 at org junit internal runner statement invokemethod evaluate invokemethod java 17 at org junit runner parentrunner 3 evaluate parentrunner java 306 at org junit runner blockjunit4classrunner 1 evaluate blockjunit4classrunner java 100 at org junit runner parentrunner runleaf parentrunner java 366 at org junit runner blockjunit4classrunner runchild blockjunit4classrunner java 103 at org junit runner blockjunit4classrunner runchild blockjunit4classrunner java 63 at org junit runner parentrunner 4 run parentrunner java 331 at org junit runner parentrunner 1 schedule parentrunner java 79 at org junit runner parentrunner runchildren parentrunner java 329 at org junit runner parentrunner access 100 parentrunner java 66 at org junit runner parentrunner 2 evaluate parentrunner java 293 at org junit runner parentrunner 3 evaluate parentrunner java 306 at org junit runner parentrunner run parentrunner java 413 at org junit runner junitcore run junitcore java 137 at org junit runner junitcore run junitcore java 115 at org junit vintage engine execution runnerexecutor execute runnerexecutor java 43 at java base java util stream foreachop foreachop ofref accept foreachop java 183 at java base java util stream referencepipeline 3 1 accept referencepipeline java 195 at java base java util iterator foreachremaine iterator java 133 at java base java util spliterators iteratorspliterator foreachremaine spliterator java 1801 at java base java util stream abstractpipeline copyinto abstractpipeline java 484 at java base java util stream abstractpipeline wrapandcopyinto abstractpipeline java 474 at java base java util stream foreachop foreachop evaluatesequential foreachop java 150 at java base java util stream foreachop foreachop ofref evaluatesequential foreachop java 173 at java base java util stream abstractpipeline evaluate abstractpipeline java 234 at java base java util stream referencepipeline foreach referencepipeline java 497 at org junit vintage engine vintagetestengine executeallchildren vintagetestengine java 82 at org junit vintage engine vintagetestengine execute vintagetestengine java 73 at org junit platform launcher core engineexecutionorchestrator execute engineexecutionorchestrator java 108 at org junit platform launcher core engineexecutionorchestrator execute engineexecutionorchestrator java 88 at org junit platform launcher core engineexecutionorchestrator lambda execute 0 engineexecutionorchestrator java 54 at org junit platform launcher core engineexecutionorchestrator withinterceptedstream engineexecutionorchestrator java 67 at org junit platform launcher core engineexecutionorchestrator execute engineexecutionorchestrator java 52 at org junit platform launcher core defaultlauncher execute defaultlauncher java 96 at org junit platform launcher core defaultlauncher execute defaultlauncher java 75 at org apache maven surefire junitplatform junitplatformprovider execute junitplatformprovider java 188 at org apache maven surefire junitplatform junitplatformprovider invokealltests junitplatformprovider java 154 at org apache maven surefire junitplatform junitplatformprovider invoke junitplatformprovider java 128 at org apache maven surefire booter forkedbooter runsuitesinprocess forkedbooter java 428 at org apache maven surefire booter forkedbooter execute forkedbooter java 162 at org apache maven surefire booter forkedbooter run forkedbooter java 562 at org apache maven surefire booter forkedbooter main forkedbooter java 548 |
openzipkinzipkin | make sense of large trace | Enhancement | description it s a trend that site be enthusiastically instrument everything in order to increase observability on what be happen or so they hope this lead to large trace be generate imagine a noisy jpa batch update on a couple of hundred record span per update or a instrumented message be publish from a producer to a horde of instrumented consumer span per poll subsequent span for the record processing and when the consumer also broadcast event then it go ad infinitum feature when present with a large trace assume this be a couple of hundred span or more zipkin ui should allow the person look at the trace to make sense of it this make sense be the topic of this issue in what way would people like to make sense of such large trace to find what they be look for some idea that have be think of query inside a trace basvanbeek can you give some specific example here what you be think of generate a summary view e g top 5 latency contributor in a trace filter a trace by span type e g toggle off the jdbc span rational large trace be a thing and as much as we would like to tell people that they be not always useful they will end up be generate and we should have a way to cope with they example scenario at our site ec one of the more large service ecosystem be constantly be confront with very large trace they be not easy to navigate or troubleshoot and the team have start to look for way to mediate this problem image not include in this issue performance improvement on client side rendering as report various time already 2496 2230 2411 1460 though the concept present here could have a positive impact for example by not show the complete trace from the start it s not the goal of this issue prior art tacigar have be do quite a few experiment on different visualization relate 2638 filter service 2920 collapse span 2496 allow to limit number of span 2296 optionally mask local intermediate span 2622 service topology view for a single trace |
openzipkinzipkin | extract storage type as jdk 1 8 | Enhancement | anuraaga notice we could consider make the storage type and in memory impl as jdk 1 8 discussion r495899024 what would need to stay in the core jar be the follow type as they be need by zipkin reporter zipkin2 call zipkin2 callback zipkin2 checkresult zipkin2 component zipkin2 endpoint zipkin2 span zipkin2 spanbytesdecoderdetector accidentally in wrong package I think zipkin2 codec bytesdecoder zipkin2 codec bytesencoder zipkin2 codec encode zipkin2 codec spanbytesdecoder zipkin2 codec spanbytesencoder zipkin2 v1 v1annotation zipkin2 v1 v1binaryannotation zipkin2 v1 v1span zipkin2 v1 v1spanconverter zipkin2 v1 v2spanconverter likely we would also need the following for those aggregate dependency if no other reason than to keep model type and codec thing together zipkin2 codec dependencylink zipkin2 codec dependencylinkbytesencoder zipkin2 codec dependencylinkbytesdecoder plus of course any internal type imply such as the codec librarie the main point be focus on which type remain in the 1 6 compatible core jar |
openzipkinzipkin | feedback need show pie chart and table instead of badge for span count | Enhancement | in 3180 we have a discussion about how to display span count at that time I create a chart with the following design but we find some problem for usability reject design so I create a new design with pie chart and table like the follow new design what do you think about this please give I your feedback cc adriancole shakuzen |
openzipkinzipkin | test docker image we would build before merge | Enhancement | I would like to enlist github action to test the docker image we would build on pr I know it will take some time but if we do it right it would likely finish before our normal integration test do it would be nice to use java to do this vs a myriad of shell script if statement which be the usual way the way I see it be an optional module similar to benchmark that a github action kick run the normal mvn install but skip normal test as we have other thing that do that in a separate project integration test prepare image use docker maven plugin run a suite of testcontainer test that initially just start the zipkin image in various config and verifie healthcheck not sure if testcontainer can automatically verify healthcheck or not later we can test other thing like post datum etc another option be to not integrate directly the thing in maven rather use a shell script to do docker build then kick off the integration test project that doesn t implicitly build the image note the building of docker image here be pre merge this process would not effect our release process in any way thought cc openzipkin core bsideup saturnism |
openzipkinzipkin | make span name count by service in the ui semantic to local root | Enhancement | feature span count by service should only return remote link base on instrument service not remote virtual one this be also the same as local root count although we don t have a local root i d yet in our model with local root instead we get the same count as use the dependency graph except this be per trace and currently exclude remote endpoint when someone see a high count they can do a search for that service name know it be a servicename not a remoteservicename implementation while I didn t implement it fix span count be use same logic as dependency linker on the spannode represent the root instead of generate link or after do that increase count only for each side correspond to the localendpoint servicename and span kind ex if link be span kind server parent foo child bar parent bar child baz parent bar child baz we count the following in this trace foo 1 bar 2 iff foo and bar be localendpoint servicename there s some tricky when the link be message this type of link be span kind consumer would be parent broker child consumerservice but a local root of the child not the parent to summarize the name to choose from the link will correspond to the localendpoint servicename in server link this end up the parent but in consumer it be the child service rational this allow we to avoid skew the brain when buggy or fine grain instrumentation result inflate the count for a specific node this be particularly important when we plot number after here be some example that skew span count profiler style tracing could show a service only hit once have an order of magnitude more span an edge service that call mysql 2000 time yet mysql be not a service buggy instrumentation that create client span when they mean to create local span example scenario will plot these number anyway do this will make the number semantically interpretable prior art dependency linker have most of this logic though it be unaware of which be instrument service and it be also unaware of the span kind of the link this mean you need a little more info to decide ex we may need to carry link kind or just implement different than dependency linker |
openzipkinzipkin | es http log body show client side healthcheck | Bug | use late with the zipkin aws image I notice the health checking for client side lb be show up in normal es http log 2020 08 24 03 27 58 804 info 1 orker epoll 2 2 c l a c l loggingclient creqid 8f26f2dc sreqid a2991add chanid e84c7112 laddr 172 17 0 2 48716 raddr search zipkin 4nynibp44iixiymakkzah3iko4 ap southeast 1 es amazonaw com 18 136 91 63 443 get request starttime 2020 08 24t03 27 58 802z 1598239678802979 length 0b duration 1699 s 1699900ns scheme none h2 name get cluster health header method get path cluster health zipkin span scheme https authority search zipkin 4nynibp44iixiymakkzah3iko4 ap southeast 1 es amazonaw com x amz date 20200824t032758z accept encode gzip deflate authorization 2020 08 24 03 27 58 877 info 1 rker epoll 2 12 c l a c l loggingclient creqid 8f26f2dc sreqid a2991add chanid e84c7112 laddr 172 17 0 2 48716 raddr search zipkin 4nynibp44iixiymakkzah3iko4 ap southeast 1 es amazonaw com 18 136 91 63 443 get response starttime 2020 08 24t03 27 58 871z 1598239678871984 length 237b duration 4774 s 4774600ns totalduration 73778 s 73778300ns header status 200 date mon 24 aug 2020 03 27 58 gmt content type application json charset utf 8 access control allow origin content encode gzip vary accept encode user agent contentpreview cluster name 993194456877 zipkin status green time out false number of nodes 1 number of datum node 1 discover master true active primary shard 0 active shard 0 relocate shard 0 initialize shard 0 unassigned shard 0 delay unassigned shard 0 number of pende task 0 number of in flight fetch 0 task max wait in queue millis 0 active shard percent as number 100 0 |
openzipkinzipkin | special case elasticsearch 7 8 composable index template | Enhancement | feature if we prefer composable template or otherwise prioritize our index template we can avoid support problem like where the template doesn t kick in due to something else match on rational composeable template be both introduce and prefer in es 7 8 which refer to template as legacy one this would introduce an issue of composeable template priority either it would have to determine automatically what value should be set or be parameterizable through env variable example scenario test would have a pre exist composable template and ours still apply thank benqs for the hint cc openzipkin elasticsearch |
openzipkinzipkin | add debugg instruction for len | Enhancement | in zipkin js we have an easy way to use a debugger for test let s have one for lens also debug |
openzipkinzipkin | text anchor center in trace view make it hard to know service name | Bug | describe the bug when a trace be open the first column show the list of service name these be set to text anchor middle with x 50 an y 50 in case where the service name do not fit in the fix width block this end up show the middle of the service name which be not ideal image step to reproduce open a trace with long service name netflix longsvcname txt expect behaviour you can visually see the starting n character of the service hover should show the full name |
openzipkinzipkin | annotation marker hide annotation when time clash | Bug | describe the bug when a span have multiple annotation with the exact same timestamp the bubble circle that show they on a linear scale seem to overlap on top of each other when you click on the bubble only the last annotation be show mask the existing of other annotation at the same point step to reproduce create a trace with a span with multiple annotation for exact same timestamp expect behaviour click on the the annotation bubble should show all annotation match the timestamp and not just the element on the top |
openzipkinzipkin | allow disable some search option | Enhancement | feature currently the zipkin ui have a standard set of trace look up criterion each of these option require an api in the backend and and index to support it some service may not have some of these due to scale constraint it would be very helpfiul if we can specify as a part of the config json which search option be support by the backend rational this help keep the ui honest on what criterion the user can actually use to find trace example scenario user want to find some trace and try to use the criterion like tag or remoteservicename which be not support by the backend prior art link to prior art more link |
openzipkinzipkin | use light shade down and allow turn off color in zipkin len | Enhancement | feature currently the span name be show in bright colored pill can we allow turn off these color additionally it would be great to use light shade instead of the bright one rational pleasing color help the user example scenario anywhere span be show prior art link to prior art more link |
openzipkinzipkin | error don t render in lens anymore | Bug | in classic I m sure api error render a warning somehow I m not sure if lens do in the past but it currently do not here s an example of a 400 which return no feedback on api error screenshot 2020 04 19 at 12 19 26 pm |
openzipkinzipkin | edit date range manually in dependency page break | Bug | put the cursor in one of the date range field press pretty much any key make it barf dependency bug |
openzipkinzipkin | allow search just by remoteservice | Enhancement | similar to 1794 feature sometimes when I m debug 504 or 503 error from a service what I want to search for be any span where that service be the destination where client status code 503 but I don t necessarily know or care which service make the call right now I can not do that because if I don t select a servicename first the dropdown for remoteservice be empty |
openzipkinzipkin | use html tag for render trace graph instead of svg | Enhancement | currently svg be use for the trace graph this be implement for render speed however normal css can not be use with svg e g text overflow this create some rendering issue like the problem of render speed can be solve by virtualization like use react virtualize so I think we should give priority to ease of draw rather than draw speed at first glance svg be very good for draw complex graph flexibly but it can be difficult to do fine thing I think we should draw the chart with normal html tag css even if it s a bit painful at first trace graph be not as complex as they can not be draw with html css |
openzipkinzipkin | nans in lensui 2 19 instead of timing | Bug | describe the bug there be a bug with trace rendering in lensui 2 19 I see nan instead of timing and there be no span display image step to reproduce upload trace |
openzipkinzipkin | add a button to easily search more span like the current one | Enhancement | feature should be an easy addition what I d like to have be a button in the expand span viewer where you can see tag annotation etc that allow you to search for more span with the same service name and span name as the current one can just be something like a redirect to the search page with pre populated service name and span name rational right now you can open a new tab and manually search for it but I think it s an easy addition that will speed up usage a bit less click require |
openzipkinzipkin | itzipkinselftrace postistrace v1 miss a span | Bug | when itzipkinselftrace postistrace v1 run in travis it consistently be miss the span for the storage call accept span it be so far not reprodicable elsewhere and for now we be ignore the test 2019 08 24 10 21 46 099 info 8830 main z s i b itzipkinselftracing start itzipkinselftrace in 0 976 second jvm run for 34 896 error test run 4 failure 0 error 1 skip 0 time elapse 11 358 s failure in zipkin2 server internal brave itzipkinselftracing error postistrace v1 time elapse 10 127 s error org awaitility core conditiontimeoutexception assertion condition define as a lambda expression in zipkin2 server internal brave itzipkinselftracing want 3 span get 2 current trace traceid 0000000000000001 i d 0000000000000002 name test trace timestamp 1566604800000 traceid 6fe88461749627eb i d 6fe88461749627eb kind server name post api v1 span timestamp 1566642106126618 duration 11372 localendpoint servicename zipkin server ipv4 172 17 0 1 remoteendpoint ipv4 127 0 0 1 port 50554 annotation timestamp 1566642106126630 value wr timestamp 1566642106137976 value ws tag http method post http path api v1 span within 10 second at zipkin2 server internal brave itzipkinselftrace awaitspan itzipkinselftrace java 119 at zipkin2 server internal brave itzipkinselftracing postistrace v1 itzipkinselftrace java 88 cause by java lang assertionerror want 3 span get 2 current trace traceid 0000000000000001 i d 0000000000000002 name test trace timestamp 1566604800000 traceid 6fe88461749627eb i d 6fe88461749627eb kind server name post api v1 span timestamp 1566642106126618 duration 11372 localendpoint servicename zipkin server ipv4 172 17 0 1 remoteendpoint ipv4 127 0 0 1 port 50554 annotation timestamp 1566642106126630 value wr timestamp 1566642106137976 value ws tag http method post http path api v1 span |
openzipkinzipkin | topology view of query result | Enhancement | this be about an alternate query screen that be topology base as oppose to table base but still use the query api and universal search paramter from the exist search screen other project like pinpoint and skywalke seem to use topology instead of query interface the result of a query would be an active image which could help people isolate the trace they want image from idea be if the datum under the trace list screen could be repurpose for a scoped aggregation where the circle could help group node by a select site tag and the fibre represent trace which you click to go to the trace view all this in javascript would require first port the dependencylinker code to javascript then do normal dependency graph to prove datum be ok then fancy thing possibly like above possibly not base on the assumption we can aggregate any way we want if do client side with small datum this be relate to as both part need a javascript dependency linker to produce the view could be use for the presentation possibly |
openzipkinzipkin | service topology view for a single trace | Enhancement | the waterfall diagram in zipkin ui be useful to understand the latency distribution for a trace however the view be not very friendly to perceive the service involve in that trace it will be useful to have another tab topology and render the topology of the service involve in the trace at netflix we trace 100 of the request that be enable for failure injection the user will find it useful to understand what service be involve in a failure inject request |
openzipkinzipkin | rocketmq as a collector | Enhancement | do zipken support rocketmq |
openzipkinzipkin | allow filter the display service in dependency graph | Enhancement | feedback from line s sre team be that when you have many service the service graph become practically unusable the ability to filter the display service would go a long way in help this user usually want to see the dependency for their service in particular or perhaps all subsequent dependency from an edge service jason from haystack mention that the vizceral library have filtering capability build in so we could easily leverage that to display a filter dependency graph |
openzipkinzipkin | improve display of long gap in trace view | Enhancement | ps for message trace there s often a long gap between producer and consumer I always wonder if we could pinch the timeline together in case like this originally post by adriancole in issuecomment 486038682 this apply to any async processing in general but when backup occur in queue processing the gap of no span can get quite big which squeeze the interesting stuff to be small if we do any cutting of blank space in the graph we would want to make sure it be visually clear to user that there be that long pause in between the span just it be hide |
openzipkinzipkin | merge span service label could be more clear | Enhancement | lens show the service involve in a display span which may be a merged span when click on the span in the ui e g the label content service and mobile gateway in the follow screenshot it may not be clear to user that these label be the service include in the merge span it be perhaps more clear in the classic ui due to the service name be preface with a service label as show below I think it would be even well if we visually or textually indicate that this be a representation of a client server rpc span perhaps as simply as put an arrow between the service from the client to the server otherwise more verbose but clearly we could indicate which service be the client and which be the server by some mean |
openzipkinzipkin | add the ability to differentiate between multiple node of a single service in the ui | Enhancement | notably datum service tracing can result in many span all classify by the same service for example cockroachdb or cassandra the trace detail screen should allow user to annotate row describe a span with the ip port possibly the same way htrace do back in the day delimit addition to the service name while it might be useful to heuristically decide if ip port should be add by default into the trace screen do so can be tricky and possibly confuse user we decide to start with just make it possible |
openzipkinzipkin | visualize annotation in len | Enhancement | span may have annotation in they for example brave flush be present when a span be abandon usually by accident we should display they the old ui use to use bubble for this but there be problem with that approach as note in 1923 |
openzipkinzipkin | site specific high cardinality key | Enhancement | 2236 add support for low cardinality site specific tag use in the universal search bar of lens site also have common key they use even if the value be high cardinality example request i d we should consider add a config json or otherwise mean to declare the site specific key the search bar can list even if the value be manual text in other word this be auto compete key but not value |
openzipkinzipkin | limit in memory span per trace | Enhancement | right now we have a setting on the in memory storage to limit total span I believe it be 500k by default timothybasanov notice that in test one could accidentally create all 500k span in a single trace which could crash the ui or lead to other issue we could relatively easily limit span count per trace in the in memory storage as it be synchronize anyway this could help with problem find in development if we do this I would recommend add a an error tag in a fake span name truncated trace or something with the amount of span drop this would make it easy to tell there be a problem and should be ok as this be a non prod storage anyway |
openzipkinzipkin | allow to limit number of span per trace for ui | Enhancement | feature additional configuration parameter to limit number of span within a single trace can potentially be apply to search result and or in memory storage rational accidentally run a big fan out job may create trace with a million span this affect user in several way zipkin ui have javascript stack overflow exception when there be too many span to render even when it do not throw browser tend to hang when deal with that much datum zipkin2 server internal zipkinqueryapiv2 writetrace have an integer overflow when result json be more than 2 gb in size user do not know which span be big but as soon as he hit it in search result it throw 500 error from the server so he can not exclude one bad trace from result non 200 error be not render in zipkin ui so user have no way of know what go wrong in memory storage behave really poorly as it iterate over all span within a single trace this disproportionally affect load testing job on a dev machine propose solution additional configuration parameter that limit total number of span send from zipkin ui within a single trace drop some datum be well then fail the whole search result ideally it should show in end user ui that some span be drop and view be incomplete this be hard to implement temporary solution after drop some span add one span back that would say error too many span within a trace have to drop some on the floor additional configuration parameter that limit total number of span within a trace for in memory storage drop some datum may be well than bring down the whole server when it either run out of memory or get stick in an endless gc loop |
openzipkinzipkin | make tooltip component show how to use in len | Enhancement | feature make tooktip component show how to use in lens to improve ux rational the new ui be significantly different than the old one I believe it be somewhat refine but it have become complicated so we need tooltip for show how to use new ui component example scenario when the user click or hover on a component show tooltip which show how to use it |
openzipkinzipkin | clearly separate autocomplete tag in lens dropdown | Enhancement | it s confuse why some tag would show up in the drop down and some wouldn t since right now everything s just mixed it d be nice if we could split the list in 2 section and move the autocomplete tag to the 2nd with ideally a title that explain that only some whitelisted tag be there and they can use the generic tag search for everything else |
openzipkinzipkin | help user share mask trace datum for debug | Enhancement | feature I want to leave the detail of how this be accomplish to discussion but the idea be that as a user you would be able to generate and share with zipkin maintainer a mask copy of your tracing datum that still reproduce an issue so that maintainer can immediately reproduce and debug the issue in short some processing that transform the original trace datum to mask trace datum in a secure way discussion question what in trace datum be generally consider sensitive what transformation can be do that will not affect reproducibility of the original issue ux how should this feature be present e g button in the ui separate endpoint script rational most zipkin site can not share their trace datum publicly as be but have trace datum that reproduce issue experience by user will enable debugging of issue to happen much fast and more accurately example scenario a zipkin user find what appear to be a bug in the way a trace be display and want to report it without leak any sensitive datum that might be in the trace with a way to share this quickly the user doesn t need to make and send mask screenshot or write description that usually result in a lot of back and forth communication before common understanding of the problem be reach prior art not aware of any prior art in this area but will update this section if some be bring to light bonus this would make it easy for we to collect additional interesting trace datum from site to use in our test datum |
openzipkinzipkin | for 2397 show key name flag when tag value of client and server differ | Enhancement | for 2397 if it s well show key name flag when tag value of client and server differ like these http path c or client nbsp nbsp nbsp nbsp nbsp nbsp content persional http path s or server nbsp nbsp nbsp nbsp content persional |
openzipkinzipkin | support hourly elasticsearch indexing | Enhancement | elasticsearch |
openzipkinzipkin | optionally mask all local intermediate span | Enhancement | when do network investigation local not remote span can interfere with intuition for example local span make trace depth less useful span count less useful it also make the trace big and hard to view adam from atlassian suggest that the ui could disable local span in the trace view possibly with a checkbox or something this would help they remove clutter when try to figure out a problem cc openzipkin ui |
openzipkinzipkin | investigate duration query concern | Question | drolando I forget if you be use cassandra or not it will help decide if the duration query by all will never work or if it be just ask for too many service name can you mention which storage you be use and if you select anything else when do duration also how many service name do you have filter by duration with service all result in a timeout every time if this will never work should we have a js check that warn against it tbh I think the current ui also have this same issue |
openzipkinzipkin | miscategorization of service label | Enhancement | the current code in the ui prefer the remote service name to the local service name this be incorrect as the local service name should always be prefer for example a client can mislabel the remote side that say there be case where we want to read the remote service name for example root span who be originate by uninstrumented client such as kinesis or leaf client span which be call uninstrumented service such as s3 or another saas dependency linker do this for example it will read the remote service name for the root and leave one way we can address this be to prefer the remote name only in these use case another way be to display both in these use case in the current ui I think we can only choose to prefer one in the next ui project lens it may be possible to present both I think haystack ui present both for all case |
openzipkinzipkin | plugin support for the ui | Enhancement | feature plugin support for zipkin ui rational people use zipkin in different way with different purpose but more importantly use it along with different tool have the possibility of include plugin for specific purpose would improve user experience all this plugin could callable hook into frontend event and they can be load into the ui by just require the js file example scenario two example a plugin that provide a dropdown with specific tag for a site could be easily plug into the ui since site vendor know what be their common tag or also this plugin can keep track of the tag use and store they in browser local storage for suggest they later correlate requestid with log this very simple could turn the request i d tag in the span detail into a link to e g kibana so you can go directly to kibana and see all log entry relate to that log |
openzipkinzipkin | visualize large swath of trace | Enhancement | from for zipkin if I have one wish it d be native visualization option for large swath of trace rational currently our only aggregate tool be the service dependency diagram while we ve discuss do single trace aggregate for a stable shape regardless of call count to redis for example we ve not discuss how to do other way of visualize large amount of trace example scenario hope to hear about this from joe so that we make the right thing prior art joe be build a multi level sankey diagram w vega in kibana in the interim please note other prior art |
openzipkinzipkin | reconsider elasticsearch indexing strategy | Enhancement | 1685 we change to use a special manual indexing of elasticsearch to work around dot in tag name without have to do terrible hack like rename dot to underscore meanwhile numerous folk have ask about have direct indexing work mostly but not always for kibana most recently narayaruna I ask xeraa about this and he have the following to say so the initial implementation be break nest object vs property with dot s like you say and be remove in 2 0 2 0 to 2 3 don t support dot in field name for 2 4 it must be explicitly enable 0 but this version add a new and work implementation if you set the minimum elasticsearch version to 5 0 dot in field name should not be a problem elastic be not support 2 x any more but I m not sure what version zipkin be target 0 enable support for dot in field name 1 so it seem we might be able to re introduce standard indexing give a version range I don t expect we can turn off 2 x but at least we could make support version easy thought cc openzipkin elasticsearch zeagord |
openzipkinzipkin | span lose with kafka and elasticsearch over capacity | Enhancement | what kind of issue be this x bug report we have zipkin 2 7 1 reading form kafka with zipkin autoconfigure collector kafka10 2 7 1 module jar and the storage be elasticsearch with high number of span we find that we be lose span we be check the metric counter zipkin collector span drop kafka in the log we get the exception 10 13 25 604 pool 1 thread 1 debug z collector kafka10 kafkacollector can not store span 1420afe62427f4ce559a0e797bb1b777 13468e1e3b2d9e26 due to illegalstateexception over capacity java lang illegalstateexception over capacity at zipkin2 elasticsearch internal client httpcall enqueue httpcall java 87 at zipkin internal v2collector record v2collector java 57 at zipkin internal collector accept collector java 79 at zipkin collector collector accept collector java 145 at zipkin collector kafka10 kafkacollectorworker run kafkacollectorworker java 84 at zipkin collector kafka10 kafkacollector lazykafkaworker 1 run kafkacollector java 196 at java util concurrent threadpoolexecutor runworker threadpoolexecutor java 1149 at java util concurrent threadpoolexecutor worker run threadpoolexecutor java 624 at java lang thread run thread java 748 we increase es max request but not fix the problem l85 l91 I think that in the case of read span form kafka if semaphore tryacquire fail it must block or the kafka reader can control it with a new call to write the span to the storage that block until there be resource available in the storage with kafka we have the buffering so we don t have the problem of the http request to write to the storage we can control easy the back pressure of the storage not read from kafka without losse any span |
openzipkinzipkin | more detail than can not decode span when return 400 | Enhancement | on jsw ask for more detail than can not decode span I think we have a bit more than that but guess we don t we should be careful to not add server load as a side effect of create a well error in other word we shouldn t return the entire input document |
openzipkinzipkin | encode json into a predefine byte array | Enhancement | narayaruna mention offline that it would be nice to encode into an exist byte array as oppose to allocate one while he didn t use this example such a change would make this drop in reusable with okio unsafecursor l1986 in do so write json can use pool buffer with ideally near zero array allocation |
openzipkinzipkin | expose swagger openapi def with the endpoint | Enhancement | from bplotnick my current use case for the swagger client be a mux scatter gather service since we re build separate zipkin cluster per zone az in this case we re go to have the ui hit a proxy that query all of the backend cluster and combine the result this also solve the different cluster case for different retention the other use case that we be go to work on soon be a processing pipeline to automatically collect and or process important trace |
openzipkinzipkin | use asciidoctor for zipkin documentation | Enhancement | zipkin documentation be comprehensive and have enough information but sometimes I feel I have to traverse to different repos to find what I need I feel it will be well to have the documentation in asciidoctor and provide a nice readable version to the user |
openzipkinzipkin | display a warning banner in case a trace be incomplete | Enhancement | an issue that we have be that since span come in from different part of our infrastructure some of they might be delay when that happen the zipkin ui will show you the span that have already be write but it win t tell you in any way that the trace be incomplete if you refresh the page a few second later you might get many more span it d be nice if the ui could show some kind of warning if the trace be incomplete there s no way to detect whether a leaf span be miss but we can easily tell whether some intermediate span be not there since the hierarchy will be break if there s even 1 span whose parent i d point to a non exist span that mean we re miss datum |
openzipkinzipkin | tag indexing whitelist | Enhancement | after switch to span v2 zipkin store all binary annotation as tag and save they for indexing if the value be less than 256 byte long this however be problematic sometimes since some tag value be very random bad for indexer and be only record to provide more information for a span to help debug e g http uri or message i d it would be good if zipkin have a configurable whitelist for tag to be index or there be something like binary annotation in the spec |
openzipkinzipkin | use new as default sort for zipkin ui | Enhancement | currently the zipkin ui use long first as default sort method while I usually find that new be use more frequently so let s vote which one do you prefer as an zipkin fan a new first b long first adriancole |
openzipkinzipkin | consider fallback in the ui when api response be too slow | Enhancement | image |
openzipkinzipkin | 100 sample aka firehose mode mostly storage affect | Enhancement | to be flesh out this would be a mode that have more impact on tracer and transport than the server that say a 100 mode likely mean less or no indexing on storage side with a short like 1hr retention there could be ui support at least fetch by i d which could query both the normal sampled storage and a firehose one or it could be consider different app these detail will be add and refresh into a revise description with a plan |
openzipkinzipkin | use elasticsearch query any yesterday s trace be very slow | Enhancement | the environment zipkin 2 3 elasticsearch 2 3 5 8 ssd node 31 g mem per node java openjdk 64 bit 1 8 0 121 os debian 8 7 3 16 39 1 query today s trace be ok servicename comqueryapi starttime 2017 11 24 08 34 endtime 2017 11 24 09 34 limit 10 result hit 666905 take 1135ms 2017 11 24 09 33 58 345 info 11768 post http 1 1 2017 11 24 09 33 58 345 info 11768 content type application json charset utf 8 2017 11 24 09 33 58 345 info 11768 content length 393 2017 11 24 09 33 58 345 info 11768 accept encode gzip 2017 11 24 09 33 58 345 info 11768 host 10 160 201 198 9200 2017 11 24 09 33 58 345 info 11768 connection keep alive 2017 11 24 09 33 58 345 info 11768 user agent okhttp 3 9 0 2017 11 24 09 33 58 345 info 11768 2017 11 24 09 33 58 345 info 11768 source false aggs traceid aggs timestamp millis min field timestamp millis term field traceid order timestamp millis desc size 10 query bool filter bool must range timestamp millis from 1511483668488 include low true include upper true to 1511487268488 term localendpoint servicename comqueryapi 2017 11 24 09 33 58 345 info 11768 end post 393 byte body 2017 11 24 09 33 59 481 info 11768 200 ok 1135ms 2017 11 24 09 33 59 481 info 11768 content type application json charset utf 8 2017 11 24 09 33 59 481 info 11768 content length 2382 2017 11 24 09 33 59 481 info 11768 2017 11 24 09 33 59 481 info 11768 take 1135 time out false shard total 12 successful 12 fail 0 hit total 666905 query yesterday s trace be very slow servicename comqueryapi starttime 2017 11 24 06 34 endtime 2017 11 24 07 34 limit 10 result hit 427373 take 16454ms 2017 11 24 09 47 47 938 info 11768 post http 1 1 2017 11 24 09 47 47 939 info 11768 content type application json charset utf 8 2017 11 24 09 47 47 939 info 11768 content length 393 2017 11 24 09 47 47 939 info 11768 accept encode gzip 2017 11 24 09 47 47 939 info 11768 host 10 160 201 198 9200 2017 11 24 09 47 47 939 info 11768 connection keep alive 2017 11 24 09 47 47 939 info 11768 user agent okhttp 3 9 0 2017 11 24 09 47 47 939 info 11768 2017 11 24 09 47 47 939 info 11768 source false aggs traceid aggs timestamp millis min field timestamp millis term field traceid order timestamp millis desc size 10 query bool filter bool must range timestamp millis from 1511476440000 include low true include upper true to 1511480040000 term localendpoint servicename comqueryapi 2017 11 24 09 47 47 939 info 11768 end post 393 byte body 2017 11 24 09 48 04 395 info 11768 200 ok 16455ms 2017 11 24 09 48 04 395 info 11768 content type application json charset utf 8 2017 11 24 09 48 04 395 info 11768 content length 2383 2017 11 24 09 48 04 395 info 11768 2017 11 24 09 48 04 395 info 11768 take 16454 time out false shard total 12 successful 12 fail 0 hit total 427373 even quere yesterday s trace in 1 second be very slow servicename comqueryapi starttime 2017 11 24 06 34 00 endtime 2017 11 24 06 34 01 limit 10 result hit 4000 take 16065ms curl v xpost header content type application json charset utf 8 header accept encode gzip header connection keep alive data source false aggs traceid aggs timestamp millis min field timestamp millis term field traceid order timestamp millis desc size 10 query bool filter bool must range timestamp millis from 1511476440000 include low true include upper true to 1511476441000 take 16065 time out false shard total 12 successful 12 fail 0 hit total 4000 use the terminate after 1 param still slow curl xpost header content type application json charset utf 8 header accept encode gzip header connection keep alive data source false aggs traceid aggs timestamp millis min field timestamp millis term field traceid order timestamp millis desc size 10 query bool filter bool must range timestamp millis from 1511476440000 include low true include upper true to 1511480040000 take 16468 time out false terminate early true shard total 12 successful 12 fail 0 hit total 12 but quere yesterday s trace in 1 second without aggs be very fast servicename comqueryapi starttime 2017 11 24 06 34 00 endtime 2017 11 24 06 34 01 result hit 4000 take 49ms curl v xpost header content type application json charset utf 8 header accept encode gzip header connection keep alive data source false query bool filter bool must range timestamp millis from 1511476440000 include low true include upper true to 1511476441000 take 49 time out false shard total 12 successful 12 fail 0 hit total 4000 |
openzipkinzipkin | display latency between cs sr ss cr annotation timing | Enhancement | the zipkin ui currently show when the above core annotation occur and the time they occur relative to the start of the trace however it would be useful to actually have the calculated latency between each of the event available if desire this could be do by have a toggle to show additional row between the exist row which outline the latency between each of the record annotation |
openzipkinzipkin | collect metric about storage layer | Enhancement | currently there be metric about ingression but not storage it would be good if some number about the storage layer e g db query success failure count response time etc be collect cc openzipkin cassandra as suggest by adrian |
openzipkinzipkin | backpressure mechanism for zipkin v2 collector api | Enhancement | we routinely get into scenario where a spike in load result in overwhelm the storage tier for example elasticsearch pool be busy and we drop a bunch of span if we have a mechanism to know a collection source such as rmq or kinesis be buffer we could choose to push back vs drop these span obviously we should be careful not to put poisonous message back etc this issue will talk about what be possible especially what be easy to support for example I ve hear from xeraa that elasticsearch beat have a backoff algorithm use to avoid overwhelming elasticsearch maybe we can look into that cc devinsba llinder shakuzen anuraaga |
openzipkinzipkin | option to display subsequent child on same line in ui | Enhancement | currently the ui display each child on a new line a b in the case of two span from the same service that do not overlap temporally a potential ui optimization would be display they on the same line a b chart by adriancole relate opentrace io support followfrom to allow the second b span to have a as a parent while still relate to b1 |
openzipkinzipkin | migrate from variable to ease upgrade | Enhancement | sometimes we need to switch storage implementation from one type to another or due to a schema incompatibility for example you may switch from mysql to cassandra or from cassandra to cassandra3 or from elasticsearch 2 x to 6 x datum migration or maintain multiple copy of storage code be one way out but due to the normal aging of trace datum another one be a fan out proxy this discuss the fan out proxy approach the proposal be to use a migrating from variable whose value be the query endpoint of an exist install ex when in a transition you leave an old service up until your migration be complete the master run the new code and query across itself and the migrating from until the latter be unset for example give an environment which you want to upgrade like so application zipkin ui query api collector elasticsearch 5 x storage component zipkin 1 29 you d install a new version and point user to it set migrating from to the old version at this point you can immediately cut write traffic to it or at some point in the future the new version will query across itself and the other migrate application application zipkin ui collector query api migrating from elasticsearch 6 x storage component query api collector zipkin 1 30 elasticsearch 5 x storage component zipkin 1 29 this eliminate the need to keep multiple copy of storage code in the same process since retention be usually day you can simply cutoff the migrating from when you be ready |
openzipkinzipkin | storage support for hazelcast | Enhancement | would storage support for hazelcast imdg imdg hazelcast org be useful if so I could submit a pr be in memory this would be similar to zipkin storage type mem except the memory storage would be on one or more external hazelcast process not part of the zipkin server itself this would give scalability and resilience without impact on performance and optionally it would be easy also to provide expiry as a configuration option |
openzipkinzipkin | storage support use influxdb interesting | Enhancement | what kind of issue be this x feature request first look at exist issue to see if the feature have be request before if you don t find anything tell we what problem you re try to solve often a solution already exist don t send pull request to implement new feature without first get our support sometimes we leave feature out on purpose to keep the project small I m interesting in help to write another storage backend for influxdb be this something that would be useful a lot of people be use influxdb to store event datum and it be really easy to setup with influx it be pretty straightforward to compare all sort of metric I m plan to use the influxdb java sdk here |
openzipkinzipkin | consider cache name write in elasticsearch | Enhancement | right now the es span consumer do stateless write to the servicespan type for example each message that contain a service and span name will try to write that name to the correspond index usually today this imply very redundant write which add load to the es cluster which isn t necessary one way to avoid this be to keep a cache of service span name point already write this could be bound so that faulty instrumentation don t end up blow up memory on the collector for example it could be limit to 2 day today and yesterday and 1000 name each implementation would be trim the map indextoservicespan which be use to write to the servicespan type if they exist in that cache we d be careful that if add to that cache we remove the entry we add if a callback fail the easy way to make such a cache without add dependency be use a hack linkedhashmap for the outer map and the keyset of a hack linkedhash map to hold the pair note very likely normal synchronization would be fine |
openzipkinzipkin | zipkin web can not search the long duration between two time | Enhancement | in zipkin web the search result be sort by timestamp and then sort on that result we usually want to find the long trace between some time but the result from search be not our want |
openzipkinzipkin | improve performance of elasticsearch query | Enhancement | right now zipkin api query translate into bucket aggregation descend by timestamp a response to a qualified query include id and timestamp and look like this take 135 time out false shard total 5 successful 5 fail 0 hit total 1000 max score 0 hit aggregation traceid agg doc count error upper bind 0 sum other doc count 0 bucket key 0068fc162f256837 doc count 1 timestamp agg value 1485173538088 value as string 1485173538088 key 00c4dda8cf431fc7 doc count 1 timestamp agg value 1485173538088 value as string 1485173538088 a follow up query look up span by trace i d in normal case only a day or two s index be use however if someone set a rather long lookback even when only a day or two match we pass the original index through this bloat the follow up request and duplicate work in elasticsearch instead we could re calculate the index from the timestamp agg in the above response |
openzipkinzipkin | how should we represent in flight span in the ui | Enhancement | currently an in flight span show up as nan as it doesn t have a duration be this ok how could we make it look well without mislead people into think an in flight span be actually complete bash curl s localhost 9411 api v1 span h content type application json d traceid 77047bf9cf04ed5d i d 77047bf9cf04ed5d name get timestamp 1484722951945904 duration 156470 annotation timestamp 1484722951945904 value sr endpoint servicename client timestamp 1484722952102374 value ss endpoint servicename client traceid 77047bf9cf04ed5d i d 40ef4e93ab3b1507 name get parentid 77047bf9cf04ed5d annotation timestamp 1484722952100852 value cs endpoint servicename client |
openzipkinzipkin | correct skew when cs cr be the parent of sr ss | Enhancement | in order to tackle 1480 on the clock skew part I think that we first need to list possible case cs cr non instrumented server cs sr ss cr with different end point rpc span with possibility of skew cs sr ss cr with same end point loopback call sr ss root span cs sr async span without response cs async span without response with non instrumented server cs cr with sr ss child the case we re try to solve here do I miss any use case |
openzipkinzipkin | ui common problem detection and highlight | Enhancement | when integrate zipkin it can happen that the instrumentation act weirdly when there be a problem in the trace example ss be after sr the trace be display badly and that s the only way to detect that there be a problem I think that there be common pattern that could be detect at runtime and we could add more and more rule with more experience the idea would be that the ui could display warning about wrong pattern to alert integrator that there might be a problem it can also be log in the collector for span problem |
openzipkinzipkin | track usability problem when rpc be split across span ids | Enhancement | some be intentionally or accidentally split rpc across different span ids for example they might intentionally fork a span I d when receive b3 header this could also happen when an intermediary fork a new span but fail to report it to zipkin in either case the client end up as a parent span as oppose to the same span as the server the goal of this issue be to support the case where a client span be a parent of the server span in the exist model this be a precondition of any work that assume single host like 939 or at least a warning of what glitch will occur until single host rpc span be support here be the main feature area to explore x ui usability how to make split span event appear in the same detail panel 1799 x clock skew correction correct skew when cs cr be the parent of sr ss x dependency linking ensure split span increment service link only once 1656 x b3 propagation ensure single host process be capture so it can be use consistently document how to handle grey box intermediary proxy that once add annotation between client and server one in the same span transparent proxy should be unimpacted as they just send header across without change they or report anything to zipkin anyway |
openzipkinzipkin | create rally benchmark for elasticsearch storage | Help Wanted | this likely should be in a separate repo but might make sense here if we create a benchmark for various storage operation we could tell if an indexing encoding or otherwise change help or hurt cc openzipkin elasticsearch help want |
openzipkinzipkin | show latency coverage for each trace in the ui | Enhancement | currently when we look at a zipkin trace in the ui we see the overall duration for the trace and then we see a list of span with individual duration however it be hard to know if the individual span add up to the actual duration of the span it would be nice to see what percentage of the total duration be cover by the span below this would be useful for a couple of use case a understand if my trace be miss some span this can be because of pipeline delay or bug in the span aggregation pipeline b since zipkin be mostly capture the network latency it will tell we how much time be spend outside of the network in the process thus quantify the white space |
openzipkinzipkin | automatically calculate and display network latency for a span in the ui | Enhancement | currently when we look at a span belong to a trace in the ui one have to mentally calculate the network latency to identify what percentage of the latency come from processing and what percentage of latency come from request processing so if a span have cs cr ss and sr the network latency would be sr cs ss cr bonus point if the network latency can be show like the chrome network latency tab add an image from chrome documentation for reference image |
openzipkinzipkin | multiple parent aka link trace | Enhancement | this be a story about trace that relate to eachother there be a number of scenario mostly message or scheduling in nature where the next task be loosely related to its predecessor here be some example where thing be relate trace a flush operation many operation in various trace be in the same flush trace a schedule task a trace operation result in a scheduled task when that task run possibly on interval can you allocate it to its origin trace a latent message you write something to a queue and consider process complete an order of magnitude or more time pass before anything proceed from there possibly nothing proceed join from multiple trace you have a batch job that s be trace it pick up a message which be late because of its origin during the trace workshop we discuss the idea of link trace in this we could add one or more link to other trace ids which either cause the current trace or be the cause of future trace you can imagine a scenario whereby create a new root span allow you to add a trace i d of its cause ex 1 trace1 know it be send off a message for offline processing it be not on the critical path of the caller the tracer generate a new root span for trace2 and add a link to trace1 trace2 be propagate in message metadata and the link of trace2 to trace1 be send out of band to zipkin in this case trace2 hasn t start yet ex 2 trace3 be an operation that need to read a message from a queue the read of a message from the queue close trace2 but the tracer s context notice trace3 be the current trace it add a link of trace2 to itself trace3 and proceed to support this span need a link array or need to create a special binary annotation of name link and value the trace i d the storage layer need to be able to search for trace by link trace i d minimally the ui need to present these link trace in the span detail view and render href accordingly |
openzipkinzipkin | postgresql support | Help Wanted | from adriancole on march 11 2016 1 19 koizumi85 ask on gitter about postgresql support spanstore jdbc take advantage of some sql syntax which might not be support in postgresql before we start down this path let s figure out some basic for example which version of postgresql be need so that we can test the configuration desire copy from original issue openzipkin zipkin java 93 |
openzipkinzipkin | add benchmark | Help Wanted | from adriancole on november 17 2015 3 21 some aspect of code need to be benchmarke as they will be run a lot here s some idea feel free to contribute for each scenario use trace of 1 5 depth and 10 200 width where span range from local to remote w up to 2 kb annotation size library microbenchmark thrift and json codec time skew adjustment integration benchmark ex server run against each spanstore impl accept second for 1 10 100 length span in json vs thrift at rest size of 10k 1 m trace service span name query latency duration query latency dependency query latency copy from original issue openzipkin zipkin java 45 |
openzipkinzipkin | document how to make a custom standalone collector | Enhancement | this involve some spring magic and isn t particularly obvious on how to do it I do it early today but I ll have to look at it tomorrow to remember what exactly be need here s what I remember so far pom modification lot of copy paste from zipkin server pom need to add the correct zipkin autoconfigure dependency to your pom so for example if you re write to cassandra add zipkin autoconfigure storage cassandra or just add all of they add spring boot configuration parser to pom add sprint boot starter to pom create the collector lot of copy paste from the kafka collector two piece configuration and stream collector it s fairly straightforward copy pasting need to add a spring autoconfigure thing to your resource create a main function to run create a main class see the main zipkin server class this class should have the sampler bean in it if you want to use environment variable to configure the app like everything else in zipkin then you need to copy paste a bunch of stuff from zipkin server yml and add that to your own app s resource finally make sure that you don t mess up the annotation compilation otherwise you ll get really weird error and it win t work at some point it might be useful to make a sample project with all of these execution be however one would normally execute a spring boot application |
openzipkinzipkin | search by dependency link | Enhancement | one thing yurishkuro mention in his presentation be that we don t support search by multiple service out of the box this be interesting because we do actually support search without a service the use case be move from a link point of view ex an edge on a dependency tree to a trace query we can investigate how search by service work on a per datastore basis as well the ui affect of such a thing |
openzipkinzipkin | indicate clearly rpc span service | Enhancement | when look at the ui the owning service of a span be display on the left hand side for example an rpc span cause by an instrumented client result in the client s service name exist this surprise some folk and doesn t look congruent with the top section where both service be mention in some case this have lead people to make local span to clarify the server side visually it seem we can make a small adjustment to fix this by visually indicate a client server span with both side the whole screen seem to make more sense for example the top count the service name and the bottom align with that count proposal |
openzipkinzipkin | performance test for integrate span collection | Enhancement | in the old repository we mention we need some work to facilitate integration benchmarke for example we need to be able to invoke reporter on demand regardless of whether that code live here or in brave while many thing need benchmarke one can reasonably argue that span collection be the more critical for example there be far more application report datum then end user of zipkin s ui or api by benchmarke collection we can help identify bug or limitation that impact zipkin s ability to perform its most basic function store span it be important that this benchmark be something that other can run as often laptop aren t representative for example in high load there be likely multiple collector shard and each may have different timeout thread pool and heap size configuration than default we could have a test that produce span against http kafka or scribe and somehow know how to analyze stat or otherwise to see how many actually arrive for example it could read the collector metric until all message be accept then look at the trace query until all span be process or timeout on timeout it could verify in a storage specific way how many span land this all be need because storage operation be async use such a tool admin can sample or throttle write to meet the performance characteristic of their integrate zipkin system for example they can set the collector sample rate accordingly or use something like zipkin zookeeper to ensure write don t exceed the capacity of the system the minimum scenario should be test reuse the same asset we do for benchmark vary on span count span message and message second it be important that these span have unique timestamp and ids and that the timestamp vary on day by use the same asset as our benchmark we can more consistently test improvement that may be library specific see 1142 961 444 |
openzipkinzipkin | zipkin ui basepath doesn t seem to work with 3 1 0 | Bug | we have upgrade to 3 1 0 we be use zipkin ui basepath to rewrite the base path for the ui but it doesn t seem to work anymore try run docker run p 9411 9411 openzipkin zipkin 3 1 0 and the ui show nicely on redirect to with docker run e zipkin ui basepath admin zipkin p 9411 9411 openzipkin zipkin 3 1 0 it still redirect to but show blank page if I try localhost 9411 admin zipkin I get 404 not find perhaps I be miss some configuration change for the upgrade |
openzipkinzipkin | spring boot admin support | Enhancement | able to register on eureka and also able to see actuator endpoint as it be fix in 3 1 0 release but zipkin be not get register on spring boot admin I have register 2 service along with zipkin on eureka eureka but on spring boot admin only 2 service be visible zipkin be not get register on admin admin kindly check thank you |
openzipkinzipkin | possbility to override the default namin strategy of span name | Question | currently the building of span name be hard code in the span builder zipkin2 span builder java public builder name nullable string name this name name null name isempty null name tolowercase locale root return this imn a distrubuted world not all use zpking as tracer for have common naming strategy it cloud be very helpgul to make that area modifyable |
openzipkinzipkin | lens i18n doesn t work anymore | Bug | if you click the language toggle nothing happen we need to make sure the language feature work both default from locale and also via language toggle before the next release cc samtv12345 if you can look into this |
openzipkinzipkin | es http logging appear break | Bug | describe the bug I be go to suggest in a different issue to use es http log body in order to know if zipkin would have a problem with preexist index which be alias I try myself first but it didn t do anything step to reproduce I apply this diff diff a docker example docker compose elasticsearch yml b docker example docker compose elasticsearch yml 37 8 37 8 service storage type elasticsearch point the zipkin at the storage backend es host elasticsearch 9200 uncomment to see request to and from elasticsearch es http log body es http log body self tracing enable true depend on storage condition service healthy then I do the follow bash cd docker example docker compose f docker compose elasticsearch yml up and then I keep click reload on which make self trace but the whole time zipkin have no output expect behaviour there should have be log in the console I m not sure when this stop work we should have have a test for this but didn t admittedly due to the packaging difference between slim and normal server former use slf4j and the other log4j2 actually get test to work might be tricky unless use invoker note that the docker compose example use slim |
openzipkinzipkin | support for a single global index | Enhancement | I have a special case where get more resource in aws be problematic because of budget constraint however we have no issue with resource for log storage hence I have an index on a very large es cluster that I can export log to but I can t get new index make since it s a company cluster they can handle expiration of datum too get funding for new aw resource be very difficult in fact I previously have a project that have aw resource repeatedly deny even though the project could save the company million in cost if it be allow to continue it take three attempt and director level override to get the 300 mo in additional aw cost approve so I ask the lead dev if we could modify the zipkin code deal with index but that isn t desirable as the company doesn t like to maintain 3rd party software so at this point I either need a global index feature fight for aws resource or find another project similar to zipkin even though we have already integrate zipkin I know it s a long shot just post to see if this be perhaps a useful feature for anyone else |
openzipkinzipkin | flakey test itzipkineureka | Bug | I m not sure if the test be a flake or if the registration that say it succeed in eureka by log anyway actually fail the read back via api fail but on the other hand it say time elapse 0 153 s I wonder if for some reason awaitility be only try this long info run zipkin2 server internal eureka itzipkineureka 11 40 38 796 warn main support postprocessorregistrationdelegate beanpostprocessorchecker postprocessorregistrationdelegate java 429 bean com linecorp armeria spring armeriabeanpostprocessorconfiguration of type com linecorp armeria spring armeriabeanpostprocessorconfiguration springcglib 0 be not eligible for get process by all beanpostprocessor for example not eligible for auto proxying the currently create beanpostprocessor armeriabeanpostprocessor be declare through a non static factory method on that class consider declare it as static instead 11 40 38 842 warn main server rejectedroutehandler rejectedroutehandler java 54 virtual host have a duplicate route health 11 40 38 912 info armeria boss http 38837 server server serverportstartlistener server java 808 serve http at 0 0 0 0 0 0 0 0 0 38837 11 40 39 086 info armeria common worker epoll 2 3 eureka eurekaupdatinglistener eurekaupdatinglistener java 175 register fv az1487 52 to eureka error test run 2 failure 1 error 0 skip 0 time elapse 11 02 s failure in zipkin2 server internal eureka itzipkineureka error zipkin2 server internal eureka itzipkineureka registersineureka time elapse 0 153 s failure org opentest4j assertionfailederror response protocol http 1 1 code 404 message url at java base java lang reflect constructor newinstancewithcaller constructor java 502 at zipkin2 server internal eureka baseitzipkineureka geteurekazipkinappasstre baseitzipkineureka java 119 at org awaitility core abstracthamcrestcondition lambda new 0 abstracthamcrestcondition java 48 at org awaitility core conditionawaiter conditionpoller call conditionawaiter java 248 at org awaitility core conditionawaiter conditionpoller call conditionawaiter java 235 at java base java util concurrent futuretask run futuretask java 317 at java base java util concurrent threadpoolexecutor runworker threadpoolexecutor java 1144 at java base java util concurrent threadpoolexecutor worker run threadpoolexecutor java 642 at java base java lang thread run thread java 1583 |
openzipkinzipkin | unable to register to an authenticated eureka server | Bug | describe the bug unable to register to an authenticated eureka server step to reproduce curl ssl bash s eureka service url java jar zipkin jar log 2024 01 18t10 53 53 319 05 30 warn 17620 main c l a s rejectedroutehandler virtual host have a duplicate route health 2024 01 18t10 53 53 420 05 30 info 17620 oss http 9411 c l a s server serve http at 0 0 0 0 0 0 0 0 9411 2024 01 18t10 53 53 741 05 30 warn 17620 worker nio 2 2 c l a s e eurekaupdatinglistener fail to register in 5cg1312171 to eureka status 401 unauthorized content |
openzipkinzipkin | java 21 java jar zipkin jar | Bug | describe the bug java lang illegalargumentexception unsupported url jar nest root zipkin jar boot inf lib zipkin lens 3 0 3 jar zipkin lens static js main d21e7047 js must start with file jar file jrt or bundle and the ui load error step to reproduce curl ssl bash s java jar zipkin jar java version openjdk version 21 0 1 2023 10 17 lts openjdk runtime environment corretto 21 0 1 12 1 build 21 0 1 12 lt openjdk 64 bit server vm corretto 21 0 1 12 1 build 21 0 1 12 lts mixed mode sharing expect behaviour ui show |
openzipkinzipkin | flakey test itelasticsearchstorage itdependenciesheavy manylink | Bug | error error error zipkin2 elasticsearch integration itelasticsearchstorage itdependenciesheavy manylink info run 1 pass error run 2 itelasticsearchstorage itdependenciesheavy itdependenciesheavy manylink 92 processdependencie 142 runtime parent datum too large datum for would be 255777200 243 9 mb which be large than the limit of 255013683 243 1 mb real usage 255254912 243 4 mb new byte reserve 522288 510 kb usage request 0 0b fielddata 0 0b in flight request 522288 510 kb model inference 0 0b eql sequence 0 0b account 1092 1 kb |
openzipkinzipkin | docker container end up with defunkt wget process | Bug | describe the bug when run the docker container for long period of time zombie procese for wget show up which can be trace back to the openzipkin zipkin container base on the code the wget seem to originate from the healthcheck as do here l48 step to reproduce keep the docker container run for a while at least on our infrastructure expect behaviour whenever wget fail for whatever reason it should be terminate gracefully and not become a zombie process |
openzipkinzipkin | add a way to install zipkin through homebrew on macos | Enhancement | feature I would like to install zipkin use homebrew use brew install zipkin rationale it would be helpful for other macos user to have a way to experiment quickly and easily with zipkin while keep it up to date with brew example scenario I get a mac I don t have docker I want to try zipkin out I would like to have a way to test it without compile prior art link to prior art more link |
openzipkinzipkin | flakey test zipkinextensiontest postspan disconnectduringbody | Bug | 3631 try to solve this by close resource but that wasn t enough error zipkin2 junit5 zipkinextensiontest postspans disconnectduringbody time elapse 0 034 s error java io ioexception unexpected end of stream on at okhttp3 internal http1 http1exchangecodec readresponseheader http1exchangecodec kt 209 at okhttp3 internal connection exchange readresponseheader exchange kt 111 at okhttp3 internal http callserverinterceptor intercept callserverinterceptor kt 95 at okhttp3 internal http realinterceptorchain proceed realinterceptorchain kt 109 at okhttp3 internal connection connectinterceptor intercept connectinterceptor kt 34 at okhttp3 internal http realinterceptorchain proceed realinterceptorchain kt 109 at okhttp3 internal cache cacheinterceptor intercept cacheinterceptor kt 95 at okhttp3 internal http realinterceptorchain proceed realinterceptorchain kt 109 at okhttp3 internal http bridgeinterceptor intercept bridgeinterceptor kt 84 at okhttp3 internal http realinterceptorchain proceed realinterceptorchain kt 109 at okhttp3 internal http retryandfollowupinterceptor intercept retryandfollowupinterceptor kt 65 at okhttp3 internal http realinterceptorchain proceed realinterceptorchain kt 109 at okhttp3 internal connection realcall getresponsewithinterceptorchain okhttp realcall kt 205 at okhttp3 internal connection realcall execute realcall kt 158 at zipkin2 junit5 zipkinextensiontest postspansv1 zipkinextensiontest java 176 at zipkin2 junit5 zipkinextensiontest assertpostspansv1success zipkinextensiontest java 184 at zipkin2 junit5 zipkinextensiontest postspans disconnectduringbody zipkinextensiontest java 131 at java base jdk internal reflect nativemethodaccessorimpl invoke0 native method at java base jdk internal reflect nativemethodaccessorimpl invoke nativemethodaccessorimpl java 62 at java base jdk internal reflect delegatingmethodaccessorimpl invoke delegatingmethodaccessorimpl java 43 at java base java lang reflect method invoke method java 566 at org junit platform commons util reflectionutil invokemethod reflectionutil java 728 at org junit jupiter engine execution methodinvocation proceed methodinvocation java 60 at org junit jupiter engine execution invocationinterceptorchain validatinginvocation proceed invocationinterceptorchain java 131 at org junit jupiter engine extension timeoutextension intercept timeoutextension java 156 at org junit jupiter engine extension timeoutextension intercepttestablemethod timeoutextension java 147 at org junit jupiter engine extension timeoutextension intercepttestmethod timeoutextension java 86 at org junit jupiter engine execution interceptingexecutableinvoker reflectiveinterceptorcall lambda ofvoidmethod 0 interceptingexecutableinvoker java 103 at org junit jupiter engine execution interceptingexecutableinvoker lambda invoke 0 interceptingexecutableinvoker java 93 at org junit jupiter engine execution invocationinterceptorchain interceptedinvocation proceed invocationinterceptorchain java 106 at org junit jupiter engine execution invocationinterceptorchain proceed invocationinterceptorchain java 64 at org junit jupiter engine execution invocationinterceptorchain chainandinvoke invocationinterceptorchain java 45 at org junit jupiter engine execution invocationinterceptorchain invoke invocationinterceptorchain java 37 at org junit jupiter engine execution interceptingexecutableinvoker invoke interceptingexecutableinvoker java 92 at org junit jupiter engine execution interceptingexecutableinvoker invoke interceptingexecutableinvoker java 86 at org junit jupiter engine descriptor testmethodtestdescriptor lambda invoketestmethod 7 testmethodtestdescriptor java 218 at org junit platform engine support hierarchical throwablecollector execute throwablecollector java 73 at org junit jupiter engine descriptor testmethodtestdescriptor invoketestmethod testmethodtestdescriptor java 214 at org junit jupiter engine descriptor testmethodtestdescriptor execute testmethodtestdescriptor java 139 at org junit jupiter engine descriptor testmethodtestdescriptor execute testmethodtestdescriptor java 69 at org junit platform engine support hierarchical nodetesttask lambda executerecursively 6 nodetesttask java 151 at org junit platform engine support hierarchical throwablecollector execute throwablecollector java 73 at org junit platform engine support hierarchical nodetesttask lambda executerecursively 8 nodetesttask java 141 at org junit platform engine support hierarchical node around node java 137 at org junit platform engine support hierarchical nodetesttask lambda executerecursively 9 nodetesttask java 139 at org junit platform engine support hierarchical throwablecollector execute throwablecollector java 73 at org junit platform engine support hierarchical nodetesttask executerecursively nodetesttask java 138 at org junit platform engine support hierarchical nodetesttask execute nodetesttask java 95 at java base java util arraylist foreach arraylist java 1541 at org junit platform engine support hierarchical samethreadhierarchicaltestexecutorservice invokeall samethreadhierarchicaltestexecutorservice java 41 at org junit platform engine support hierarchical nodetesttask lambda executerecursively 6 nodetesttask java 155 at org junit platform engine support hierarchical throwablecollector execute throwablecollector java 73 at org junit platform engine support hierarchical nodetesttask lambda executerecursively 8 nodetesttask java 141 at org junit platform engine support hierarchical node around node java 137 at org junit platform engine support hierarchical nodetesttask lambda executerecursively 9 nodetesttask java 139 at org junit platform engine support hierarchical throwablecollector execute throwablecollector java 73 at org junit platform engine support hierarchical nodetesttask executerecursively nodetesttask java 138 at org junit platform engine support hierarchical nodetesttask execute nodetesttask java 95 at java base java util arraylist foreach arraylist java 1541 at org junit platform engine support hierarchical samethreadhierarchicaltestexecutorservice invokeall samethreadhierarchicaltestexecutorservice java 41 at org junit platform engine support hierarchical nodetesttask lambda executerecursively 6 nodetesttask java 155 at org junit platform engine support hierarchical throwablecollector execute throwablecollector java 73 at org junit platform engine support hierarchical nodetesttask lambda executerecursively 8 nodetesttask java 141 at org junit platform engine support hierarchical node around node java 137 at org junit platform engine support hierarchical nodetesttask lambda executerecursively 9 nodetesttask java 139 at org junit platform engine support hierarchical throwablecollector execute throwablecollector java 73 at org junit platform engine support hierarchical nodetesttask executerecursively nodetesttask java 138 at org junit platform engine support hierarchical nodetesttask execute nodetesttask java 95 at org junit platform engine support hierarchical samethreadhierarchicaltestexecutorservice submit samethreadhierarchicaltestexecutorservice java 35 at org junit platform engine support hierarchical hierarchicaltestexecutor execute hierarchicaltestexecutor java 57 at org junit platform engine support hierarchical hierarchicaltestengine execute hierarchicaltestengine java 54 at org junit platform launcher core engineexecutionorchestrator execute engineexecutionorchestrator java 147 at org junit platform launcher core engineexecutionorchestrator execute engineexecutionorchestrator java 127 at org junit platform launcher core engineexecutionorchestrator execute engineexecutionorchestrator java 90 at org junit platform launcher core engineexecutionorchestrator lambda execute 0 engineexecutionorchestrator java 55 at org junit platform launcher core engineexecutionorchestrator withinterceptedstream engineexecutionorchestrator java 102 at org junit platform launcher core engineexecutionorchestrator execute engineexecutionorchestrator java 54 at org junit platform launcher core defaultlauncher execute defaultlauncher java 114 at org junit platform launcher core defaultlauncher execute defaultlauncher java 86 at org junit platform launcher core defaultlaunchersession delegatinglauncher execute defaultlaunchersession java 86 at org apache maven surefire junitplatform lazylauncher execute lazylauncher java 56 at org apache maven surefire junitplatform junitplatformprovider execute junitplatformprovider java 184 at org apache maven surefire junitplatform junitplatformprovider invokealltests junitplatformprovider java 148 at org apache maven surefire junitplatform junitplatformprovider invoke junitplatformprovider java 122 at org apache maven surefire booter forkedbooter runsuitesinprocess forkedbooter java 385 at org apache maven surefire booter forkedbooter execute forkedbooter java 162 at org apache maven surefire booter forkedbooter run forkedbooter java 507 at org apache maven surefire booter forkedbooter main forkedbooter java 495 cause by java io eofexception n not find limit 0 content at okio realbufferedsource readutf8linestrict realbufferedsource kt 332 at okhttp3 internal http1 headersreader readline headersreader kt 29 at okhttp3 internal http1 http1exchangecodec readresponseheader http1exchangecodec kt 179 87 more |
openzipkinzipkin | flakey test itactivemqsender | Bug | itactivemqsender loop try to connect to the vm transport in ci sometimes here s an example from the jdk11 job 2023 12 13 04 59 18 131 activemq task 2 warn org apache activemq transport failover failovertransport fail to connect to vm embed broker create false after 30 attempt s with broker name embed broker do not exist continue to retry we should probably switch to a docker test image approach as we already have problem with this test library as it be pin to junit4 |
openzipkinzipkin | ci helm release process be independently versione and release vs zipkin | Bug | the helm release process isn t document in build bin readme but look at tag it seem to require manual and separate automation from our normal release which also publish javadoc and docker we should not have separate version and instead be like docker where we use the version of zipkin as the primary version version update in helm should be seamless on release an alternative could be to move helm to another repo where the independent versioning could make sense in the release history however I think it be well to just align the process |
openzipkinzipkin | cve 2023 46604 support for zipkin server | Bug | hi I be use zipkin server this look like a server side cve once your server get upgrade you should be good I find the above description in this issue be there any plan to raise the activemq client version to 5 16 7 on zipkin server if not we would be glad to know the procedure if it be possible to update this file here sorry if there be any misinterpretation because I write this text in deepl |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.