text stringlengths 0 1.49k |
|---|
**Brian Ketelsen:** Yes, times always get hard a lot, especially when you're on the bleeding edge. There's craziness, for sure. |
**Erik St. Martin:** Scott Mansfield is trolling us in the GoTime FM channel. He said, "Gotta fill that container spot on your buzzword bingo card." \[laughs\] |
**Brian Ketelsen:** There are definitely benefits to containers, especially in the development environment. There are certainly times where doing things on a Mac is just nowhere near as easy as doing it in Linux, so I'm all over the development environment Linux thing. But I agree, especially with Go, that there is les... |
**Erik St. Martin:** \[12:12\] I mean, you get process isolation and stuff like that too, where you can control using CGroups how much resources it can consume, namespaces - you can kind of isolate processes from each other, if your project has those sorts of needs. |
**Charity Majors:** Yes, absolutely, but you're adding complexity, so that's the tradeoff. I don't know if you've seen... JessiTron has this great blog post about where it's okay to experiment and add complexity... Because you don't wanna take people's playfulness away from them. We all do tech because we love it, and ... |
**Erik St. Martin:** There was a post I read, I wish I remembered it now... It was basically saying you have a budget for new tech and bleeding edge, and you get one. |
**Charity Majors:** Yeah, that was Dan McKinley. You get three innovation tokens as a startup. |
**Erik St. Martin:** Yes, and different things are worth different amounts, or something like that. I remember, if you picked some hot new key-value store, that's your one; everything else better be proven technologies. I think that's a smart move, because if everything is new and bleeding edge, you're gonna be fightin... |
So let's talk a little bit about what's going on at Honeycomb, what all of you are building there,, and kind of how Go is proving itself. |
**Charity Majors:** Yeah, so what we are building is a tool for debugging complex system. It's kind of like -- you can think of GDB for systems, sort of... Or you can think of it like your IDEs for your code. A lot of people will throw around these terms like "predictive analytics" and "machine learning", and... Come o... |
**Erik St. Martin:** Now, for anybody who's not familiar, do you wanna give a brief explanation of what Scuba is? |
**Charity Majors:** Yeah, absolutely. I'll tell you what Honeycomb is, because it's less Facebooky. We accept JSON at the edge and arbitrarily wide set up key-value pairs. We aggregate in real-time on all of those dimensions at once, so there's no indexes. You don't have to pick the three or four or five things that yo... |
\[16:10\] We pre-compute things like percentile buckets, max and min are always there... Having that raw, original row, like getting to deal with events while you're debugging is mind-blowing. It's so impossible to go back to aggregates or rollups or ticks or counters after that. On the backend we had to write our own ... |
**Brian Ketelsen:** You got to rewrite Cassandra? |
**Charity Majors:** No, not really... It's not even close, actually. They're like distant cousins, sure. But it's not a database. People give me shit all the time about writing a database after spending my entire life telling people not to write databases; we didn't, we did not write a database. There was hardly any qu... |
When systems get sufficiently complex, they outstrip your ability to predict what is going to break. And I think a lot of us are hitting that threshold faster and sooner than ever before, because there are so many trends that are pushing this level of complexity - everything from schedulers and containers to \[unintell... |
It's really important to us to be able to debug your database. I don't know how people DBA without this kind of thing, and the answer is that they don't. They just know how to look for slow queries, but that's very often not actually the problem. For instance, people are like "Oh, my database is getting slow, and I loo... |
The problem - with our tool what you can do is add up all of the time that each lock is being held by each lock query, and there it is. It's just so easy to deviate once you can see these things. I get down to the weeds of the databases I guess if Cristina is here it's way more interesting Go stuff to say, but we're us... |
**Brian Ketelsen:** So when you're putting together Honeycomb, the idea is that you capture and collect all of the data from all of the different pieces in these distributed systems that we're building today... |
**Charity Majors:** Correct. |
**Brian Ketelsen:** ...and then use that to correlate and discover problems or maybe predict problems before they come - that's the idea behind Honeycomb, that it's kind of the central store for all of the information, all events that are going on and help you debug? |
**Charity Majors:** \[20:07\] Yeah, it's so easy to just capture events from everywhere. Why do we have a different thing that software engineers are supposed to look at - a different thing than DBAs look at... All this does is it creates these barriers of language and of tooling and of use of reality. You don't have a... |
People spend so much time talking about what their tools says, versus what somebody else's tool says, and you're not actually even talking about the problems here. |
**Erik St. Martin:** It's the network. It's always the network. Blame those guys, |
**Charity Majors:** It's always the network's fault. Well, what if you could see the network, too? Then you capture it in the form of JSON? Cool, then that might be able to tell you something. |
**Brian Ketelsen:** That's awesome. So you said the UI is written in Go; you actually shipped a web UI in Go and nobody died? |
**Charity Majors:** Nope, nobody died. |
**Brian Ketelsen:** Wow. |
**Charity Majors:** Now, we don't have to ship that much... Christine, my co-founder is amazing at holding the line against bad UX. There are so many features that we've actually implemented in the backend, and she won't let us put it up, because she's like, "Has this been designed? Have we thought about this? No, I wi... |
**Brian Ketelsen:** You do, you gotta have somebody to gate all of those things. Erik and I have that going really well together. I have crazy ideas and he always tells me no. It works well. |
**Charity Majors:** It's fine to move fast, but when it comes to UI, it is so hard to untangle good design and get back to a place that's sane. |
**Erik St. Martin:** And every time you change it you end up upsetting some subset of the people using it. People don't even have to use something; if you take it away from them, they feel lost. "What if I needed it?" |
**Charity Majors:** Yeah... "I knew how to do that. I know I'm gonna get paged in the middle of the night someday and I'm gonna need that button." |
**Erik St. Martin:** So I think it is about time for our first sponsored break. Our first sponsor today is Toptal. |
**Break:** \[22:07\] |
**Erik St. Martin:** And we are back, talking with Charity Majors from Honeycomb. We were just talking about the UI... So that's written in Go? |
**Charity Majors:** Yes. |
**Erik St. Martin:** Is this like GopherJS? Or you just needed the frontend layer. |
**Charity Majors:** Yes, the query -- we are using JavaScript for the actual presentation layer. |
**Erik St. Martin:** Okay. |
**Brian Ketelsen:** Aww... |
**Charity Majors:** \[laughs\] Sorry, kids. |
**Erik St. Martin:** Brian would have done a backflip. |
**Brian Ketelsen:** I'm crushed. |
**Charity Majors:** We're apologetic |
**Erik St. Martin:** Is that React? |
**Charity Majors:** Yeah. |
**Erik St. Martin:** Okay. |
**Charity Majors:** You've gotta keep frontend engineers happy. |
**Erik St. Martin:** So that seems to be the thing people are doing now with a lot of web apps, in Go anyways - a JSON HTTP interface from Go, and doing React and stuff on the frontend. |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.