text stringlengths 0 2.35k |
|---|
[866.10 --> 869.32] So, make sure you add those checks in place. |
[869.52 --> 872.98] So, when it fails, it can smoothly recover and all of those. |
[873.80 --> 874.04] Yeah. |
[874.36 --> 874.56] Yeah. |
[874.56 --> 882.40] And I know some companies that have that as part of the proper sort of testing approach is they'll literally things will break on purpose. |
[882.60 --> 885.32] And, you know, it's a first class concern that they have. |
[885.32 --> 890.20] And it is that thing of, yeah, don't, you know, of course, like, I don't know. |
[890.26 --> 894.40] Is it just ego that people think I'm so good, I'll write this, it's going to be great? |
[894.60 --> 895.22] What's going on? |
[895.32 --> 896.20] It can't be. |
[896.52 --> 897.68] Honestly, it can't be, right? |
[897.70 --> 901.10] Like, we've all known and experienced it enough. |
[901.34 --> 901.60] Yeah. |
[901.72 --> 902.16] Do you think? |
[902.16 --> 902.68] I don't know. |
[902.72 --> 911.40] But the thing is, when I'm writing code and it doesn't work, it's shocking how quickly I'm like, there's something wrong with the processor. |
[912.32 --> 914.92] The processor is not working. |
[915.06 --> 916.12] Or physics has changed. |
[916.56 --> 919.58] That's why I'll go to physics has changed before it's my fault. |
[920.06 --> 922.68] But it turns out I just did a capital letter where I shouldn't have. |
[922.68 --> 929.64] I think, like, I've been in teams where they do pairing and, like, those mobbing sessions and stuff. |
[929.88 --> 938.20] And, like, they have kind of helped in, like, sense-checking people's egos a bit and be like, I'm not the best. |
[939.00 --> 944.70] And when two people talk about it, I think it does help think, okay, there is a reality that we live in. |
[944.86 --> 946.84] And this is what you need to consider. |
[946.84 --> 962.72] Is there anything that you would, like, even in that, like you were saying before, that progression between, like, being in, you know, manual testing and then QA and then moving to SRE, was there any, was there, like, a moment where it clicked, where people started, I mean, they just started incorp... |
[962.76 --> 971.88] Do you see the same progression happening in observability to where there'll be some moment and some kind of click where it just becomes part of it as opposed to the separate thing that happens afterwards? |
[971.88 --> 974.34] I have seen it work in some teams. |
[974.54 --> 978.02] And, like, a lot of teams I've worked in are all autonomous teams. |
[978.16 --> 982.68] So, they can basically build however they want using whatever technologies they want. |
[983.12 --> 995.58] What is often have teams like that is having some sort of guardrails, which actually says, and, like, also being aware that not all applications need the same level of checks and monitoring and all of these. |
[995.58 --> 999.36] So, being aware that, okay, there is a level of criticality of my app. |
[999.54 --> 1003.44] And if it is a highly critical app, then let me put all of the things in. |
[1003.84 --> 1010.60] And if it is a less critical app, in that case, you would just have, like, maybe just a simple health check. |
[1010.70 --> 1011.60] That would be good enough. |
[1011.68 --> 1013.50] You don't need to go all board. |
[1013.50 --> 1022.90] And one thing that we had when I was working for FT was you always have at least a basic check on all your apps. |
[1023.24 --> 1028.56] Otherwise, like, we used to get this service operability score for our applications. |
[1028.88 --> 1033.72] And the score used to go down as in when we didn't have some of these things in place. |
[1034.08 --> 1040.96] And that was, like, a nice measure where people thought about it from the beginning rather than as an afterthought about some of these. |
[1040.96 --> 1049.94] But what could happen in this kind of scenarios is people go all in and they just say, oh, I'm going to monitor everything, have all my logs go in. |
[1050.10 --> 1052.46] Like, you don't need to go all board on this. |
[1052.56 --> 1055.14] There's a limit to how much you need to monitor as well. |
[1055.30 --> 1063.00] And understanding the criticality of your app and then building your observability around that is probably something that teams should think about. |
[1063.00 --> 1073.86] How would you, if a team was listening to this, right, and they were trying to understand the criticality of the app and make decisions around it, like, how would you, if you had them sitting in the room, how would you explain it and say, here, start here, do this? |
[1073.86 --> 1076.56] I think it depends on the business criticality. |
[1076.84 --> 1086.24] And if it is a highly business critical application, which means if it went down for, say, more than 15 minutes, then we wouldn't be in business. |
[1086.24 --> 1100.46] If it's that kind of app, then you need to have your alerting in place, monitoring, like, the right level of logging in place, which actually gives us any of the audit records that actually show us what's happened with the applications. |
[1101.32 --> 1104.10] And then any sort of health check. |
[1104.16 --> 1108.08] So there's probably, like, two levels of monitoring that we should think about. |
[1108.18 --> 1109.76] One is the application level monitoring. |
[1109.76 --> 1111.96] And then there is the system level monitoring. |
[1112.16 --> 1120.04] So being able to figure out where the problem is soon enough is something very critical when it's a 15 minutes recovery thing. |
[1120.30 --> 1132.34] But if it is an application that's less critical, then maybe just having the application level monitoring is good enough where you could take longer to actually investigate, look into the locks and actually figure out where the problem area is and stuff. |
[1132.34 --> 1137.66] So I would suggest teams to think about, like, how critical their app is. |
[1137.76 --> 1143.50] And that is something the business should, like, help them with, not something that the team just decides, oh, this is the most critical thing. |
[1143.92 --> 1155.60] And once you know the business criticality of something, then it is coming up with some sort of check saying if it is a highly critical system, then we do both application as well as system monitoring. |
[1156.08 --> 1159.12] Otherwise, just one of them based on your use cases and stuff. |
[1159.12 --> 1166.40] And, like, in the past, I've spoken about, like, the use method and red method that we could use for these kind of things. |
[1166.68 --> 1171.02] Like, I prefer use and red method over the Google's four signals. |
[1171.18 --> 1176.20] It depends on what your team's needs are and what fits into your use cases. |
[1176.84 --> 1183.22] So you would use a red method, which is rate, error, and duration for every single application that you build. |
[1183.22 --> 1188.72] And it's very easy to see that in a microservice world where you have different kinds of applications. |
[1189.38 --> 1192.84] You have the same three parameters that you're measuring across all of them. |
[1192.96 --> 1199.62] So it actually helps the team analyze irrespective of if that belongs to your team or any other team. |
[1199.70 --> 1201.10] You just know where the problem is. |
[1201.46 --> 1203.82] And the same with systems side of things. |
[1203.82 --> 1208.40] You would go with, like, the use method, which is utilization, saturation, and errors. |
[1208.86 --> 1213.86] And you would do this for the CPU, disk, or network, and all of those different areas. |
[1214.06 --> 1216.82] And you basically know where the problem is. |
[1216.86 --> 1218.06] And it's easy to find out. |
[1218.40 --> 1220.24] I would say it is hard. |
[1220.44 --> 1221.68] It takes time. |
[1221.98 --> 1226.96] So invest based on how much returns you would get on these when you put these checks in. |
[1226.96 --> 1233.96] So that is something the team should be mindful about when they are investing in monitoring or, like, learning and stuff. |
[1234.88 --> 1239.90] Is the primary counterbalance, in your mind, the effort that it takes to keep this monitored well? |
[1240.00 --> 1240.88] Or is it also cost? |
[1241.06 --> 1243.88] Do you think about the cost to operate or the back end? |
[1244.52 --> 1245.74] It is the cost. |
[1246.00 --> 1248.72] And at the end of the day, it should be the cost to the business. |
[1248.72 --> 1253.22] As in, how much does having the system down cost us? |
[1253.70 --> 1260.80] And you basically work backwards from there, saying, if this was down for 15 minutes, it would cost the business so much. |
[1260.88 --> 1271.34] Which means we, as a team, should be investing more time in actually getting the right amount of measures so we can solve the problem or narrow down the problem quickly. |
[1271.34 --> 1278.98] And, like, I would always focus on the business value rather than the team's individual product value and stuff. |
[1279.44 --> 1287.52] But, yeah, it depends on, like, if you were an internal system, like, in one of the teams I was in, we were building monitoring tools for other teams. |
[1287.72 --> 1291.68] So we don't have real business value as such as our team. |
[1291.88 --> 1296.76] But we were supporting teams that had, like, really high value systems. |
[1296.76 --> 1304.68] So that kind of meant that we had to think about the application level as well as system level monitoring on our systems and stuff. |
[1308.98 --> 1312.98] This episode is brought to you by our friends at FireHydrant. |
[1313.26 --> 1316.04] FireHydrant is the reliability platform for every developer. |
[1316.46 --> 1320.24] Incidents, they impact everyone, not just SREs. |
[1320.40 --> 1325.98] They give teams the tools to maintain service catalogs, respond to incidents, communicate through status pages, |
[1325.98 --> 1328.14] and learn with retrospectives. |
[1328.50 --> 1333.90] What would normally be manual error-prone tasks across the entire spectrum are responding to an incident. |
[1334.18 --> 1337.36] They can all be automated in every way with FireHydrant. |
[1337.66 --> 1342.74] They have incident tooling to manage incidents of any type with any severity with consistency. |
[1343.28 --> 1346.42] Declare and mitigate incidents all from inside Slack. |
[1346.80 --> 1349.82] Service catalogs allow service owners to improve operational maturity |
[1349.82 --> 1353.12] and document all your deploys in your service catalog. |
[1353.12 --> 1359.36] Incident analytics allow you to extract meaningful insights about your reliability over any facet of your incident |
[1359.36 --> 1361.08] or the people who respond to them. |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.