text stringlengths 0 1.49k |
|---|
• Kube-Lego project by JetStack for Kubernetes Ingress controllers |
• Let's Encrypt certifications through Kubernetes |
• ACME TLS certificates fetched automatically |
• Profit motive vs. security in managing vulnerabilities |
• Production setup and limitations of the show |
• Show format and behind-the-scenes explanation |
• Shoutouts to listeners, sponsors, and guests |
**Erik St. Martin:** Welcome back everybody for another episode of GoTime. Today's episode is number \#35 and our sponsors for today are Toptal and Compose. |
Today on the show we have myself, Erik St. Martin, Carlisia Pinto is also here. Say hello, Carlisia... |
**Carlisia Thompson:** Hi, everyone. |
**Erik St. Martin:** And Brian Ketelsen... |
**Brian Ketelsen:** Hey! |
**Erik St. Martin:** And today's special guest is a co-founder of Honeycomb.io, formerly of Parse, but I don't wanna give away too many details... I'll let her introduce herself and a little bit of her Go background. Please welcome Charity Majors. |
**Charity Majors:** Hey, thanks so much for having me. |
**Erik St. Martin:** So you wanna give everybody maybe a little bit of your background, how you got into Go and what you're using it for? |
**Charity Majors:** Yeah, totally. I started using Go at Parse; we wrote the platform originally in Ruby, and it just hit a wall. We wrote the entire thing in the course of two years in Golang. It turned out it was much harder to rewrite than to write, especially with mobile clients, because you know, everybody ships t... |
**Brian Ketelsen:** Good choice. |
**Erik St. Martin:** That's a good point, designing APIs for mobile devices... Because you can't control when people upgrade, and some people never update. |
**Charity Majors:** They never update, and we had a million of them. Planned obsolescence is definitely a thing that we could do and we didn't, because we would just lose a significant faction of our users. |
**Erik St. Martin:** Was this performance bottlenecks you guys were running up against, or...? |
**Charity Majors:** Well, that was the one that we had first... We had a bunch of unicorns; you know, the Ruby web server... Ruby isn't exactly thread-safe so we've got this fixed pool of unicorns, and we had first one database and it was pretty okay, but by the time we had 25 replica sets behind it and MySQL and Cassa... |
**Brian Ketelsen:** That is such a familiar story. I've never heard that before. \[laughter\] And I gotta be honest, I don't miss unicorns, I really don't. |
**Charity Majors:** You know, the unicorn thing... I'm very much associated with a unicorn in most people's minds, and I think it's because it's fun and sparkles and magic, and it's Unicorn because that web server was just the pain of my existence for years. |
**Brian Ketelsen:** Ugliness. |
**Erik St. Martin:** Is everybody using Puma now? I think that's the thing now. |
**Charity Majors:** \[04:02\] Using what? |
**Brian Ketelsen:** Everybody's using Go now. Nobody's using Puma anymore, they're using Go. |
**Erik St. Martin:** The people who have not yet converted... |
**Charity Majors:** You know, Ruby's fine... I mean, Unicorn is fine, for what it does. If you're writing a web app - fine; it has a lot of helpers. Would it have taken us longer to prototype and write Parse the first time? No question about it, and you have to find slightly better engineers also. |
So I don't even know that I would have made a different decision that they did in the beginning... Because remember, when you're a startup, it is never clear that you're gonna succeed, and Ruby helped this get going really fast. And Go didn't really exist outside Google, remember? We didn't have this pool of Go program... |
**Erik St. Martin:** Here's a question, because the Parse rewrite was pretty early on, so I'm assuming everybody grew the talent internally; there was nobody on the team who was already familiar with Go... You were just kind of learning on the fly. |
**Charity Majors:** Correct, but we evaluated all of these statically typed languages and we argued about it for like six months, and the reason that we chose Go -- it came down to C\# and Go, really... And the reason that we chose Go was as a recruiting thing. Because it became really clear that people wanted to work ... |
**Brian Ketelsen:** That's interesting. I think that's the first time I've heard that. I've definitely heard of people rewriting in Go for its technical merits, but for its HR recruiting merits - it's not one that I've heard yet before. That's pretty awesome. |
**Charity Majors:** It was definitely the clincher. |
**Carlisia Thompson:** I know you're still small, but I know you're probably either hiring here and there or you're testing the waters... How do you see the pool of developers for Go, and how do you even approach it? If you had an opening right now, would you be looking for somebody with experience in Go, or somebody j... |
**Charity Majors:** No, I don't care what languages people know, honestly... I have huge faith in people's ability and capacity to learn. Actually, I wouldn't want to hire someone who I didn't think could learn whatever we needed. I hate it when people do that sort of thing... "Oh, I see you have experience in Blah." N... |
**Brian Ketelsen:** Couldn't agree more. |
**Erik St. Martin:** Having a good foundation and the passion and interest to learn is more important sometimes. |
**Charity Majors:** Totally. |
**Erik St. Martin:** Brian and I have talked about this... I like hiring people for who they're gonna be tomorrow, versus who they are today. Because some people are smart, and they'll be that same person tomorrow, whereas the other guy or girl will be ridiculously smarter a year from now and surpass the people you alr... |
**Brian Ketelsen:** So tell us about how you're putting Go in the stack at Honeycomb. What kind of interesting challenges have you come up with there? |
**Charity Majors:** Boy, all of our interesting challenges at this stage are around driving user adoption. I just can't even claim otherwise... It would be ridiculous. One thing that I will call out though is that people keep asking me when I'm gonna use containers, and I'm like "Why? I'm running Go." I often think tha... |
It is very much the same at Honeycomb. You've got a tiny little statically linked binary, just copy it around places, it's great. |
**Erik St. Martin:** \[08:17\] Yeah, I mean... I guess the motivation for containers could be orchestration, too. If you have a large cluster... |
**Charity Majors:** Yeah, if you have enough to justify schedulers, yes. But almost nobody does, and it's self-inflicted damage when they try. |
**Erik St. Martin:** I was talking with a group of people at KubeCon this year, which is the Kubernetes conference, and somebody was asking, "When should they scale to containers or orchestration platforms", and my first pass at that was "When you can't name all of your machines by hostname." Saying like, "Chicago Web ... |
**Charity Majors:** Yeah. |
**Brian Ketelsen:** So my one Node Kubernetes cluster is overkill? |
**Erik St. Martin:** \[laughs\] |
**Charity Majors:** Not for you. You have fun with that! \[laughter\] |
**Erik St. Martin:** Every technology you adopt comes with overhead too, right? Teams need to learn and understand, and they need to maintain it, they need to know what the failure scenarios of those things are, how to debug it... So it's just one of those things, it's added overhead if you don't have a team or infrast... |
**Charity Majors:** Totally. It's another layer of abstraction, and I like that policy about not naming your host. I haven't done that in like decades. But yeah, orders of magnitude... And it's also a function of the complexity, the number of host types that you have. I do like the way that we're going, but a lot of th... |
**Erik St. Martin:** Yeah, I think that with all of that stuff it's an early technology, which means it's not fully flushed out yet, which means that if you're struggling and you can gain from it and it's something that you can build off of and save yourself problems and engineering time, then awesome. But if you're no... |
**Charity Majors:** Oh, for sure. Yeah, there's a lot of really good stuff there. The bones are great, and it's exciting to get to play with some of these toys. It's just that we - exactly as you said - we'd have to know that they are toys for most of us. And if you don't have the drive to go all in on it when times ge... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.