text stringlengths 0 2.29k |
|---|
**Florian Forster:** Exactly. I mean, if you asked me whether I like Go generics or something like that, I need to resort to the answer "I have not experienced them yet, because I haven't had time to look into it." |
**Gerhard Lazu:** Okay... |
**Florian Forster:** My engineers will tell you a different story, but I'm not opinionated on that end. |
**Gerhard Lazu:** So what does Citadel do? And maybe I'm mispronouncing it, but I'm sure that you will give us the official pronunciation. |
**Florian Forster:** I mean, we call it Citadel as well, so it's totally fine. And the logo kind of tries to reflect the word origin, because it comes from the French way of building fortresses, with the star design... The place where you commonly see this is in Copenhagen, for example; there's still like the fortress ... |
\[09:45\] What we basically try to do is we want to bring the greatness of zero, like a classic closed source proprietary cloud service, with the greatness of key cloaks, run yourself capabilities, and we want to combine them into one nice, tidy package, so that basically everybody who has at least a heart in engineeri... |
**Gerhard Lazu:** Okay. Okay. So there's one word that comes to mind when you see that. Actually two. CPU and hashing. \[laughter\] So from all that description, all I'm thinking is "I hope your CPU is fast, and your hashes are even faster." But don't skimp on the cycles. Right? Like, make sure you do enough iterations... |
**Florian Forster:** Yeah, I mean, it comes down to hashing. We rely on Bcrypt normally, and yes, we need many CPU cycles... But also, there's like a second thought, that if you do signing tokens - so that's also quite exhaustive for CPUs nowadays. And so yes, we use a lot of CPU to make that happen. |
But actually, the stress is somewhat alleviated down to the future, because the pass key concept and \[unintelligible 00:11:30.03\] in general, since that relies on public-private key cryptography, it reduces quiet well on our end the amount of CPU we need to use, because RSA signature verification normally takes aroun... |
**Gerhard Lazu:** Yeah, yeah. Okay. Do you use GPUs for any of this? |
**Florian Forster:** No. At the moment, no. I mean, it comes down to a certain degree that our cloud provider does not really allow for that... And on-premise environments also are oftentimes really restricted in the kind of GPU they have around... And so we try and avoid for Citadel too much of integration and depth t... |
**Gerhard Lazu:** So most of the workloads that you run, I imagine they're CPU-bound, because especially of the hashing. But I also think network is involved. So you need to have a good network. I'm gonna say good network - not high throughput, but low-latency network. Can you predict these workloads? I mean, people si... |
**Florian Forster:** Yeah, we have a common pattern called double-dip traffic. So we often see like traffic starting like seven o'clock, until twelve, and then from one to around seven in the evening. So that's the major traffic phase. But if you spread that across the globe, you see certain regions ingest traffic duri... |
\[13:45\] And yeah, having a network that can be easily and fast scalable, and really provides low latency to our customers, because they use our API, our login page, is kind of important to the service quality. And so we internally set our goal to "Okay, let's try and keep latency below 250 milliseconds for 80% of the... |
**Gerhard Lazu:** Yeah, that's right. Especially if you're in like SAML, or something like that, where you yourselves need to talk to another provider, which itself may have variable latency at specific periods of time... And then you can only guarantee what you get, basically. You can't make (I don't know) GitHub logi... |
**Florian Forster:** Yes. It's bound to external services. And in our case, for the storage layer we mainly rely on CockroachDB, since they can move around and shift data closer to data centers that they are actually being used. But still, the data is the most heavy thing right there if you want to reduce latency. |
**Gerhard Lazu:** Okay. What does your tech stack look like? Like, what do you have? What is Citadel running on? |
**Florian Forster:** Basically, we use internally Cloud Run for our offering, so Google Cloud Run, because we can easily scale during traffic peaks, as well as CPU peaks, because Google will scale Cloud Run containers as well based on CPU usage. That fits our narrative quite well. And also, the ability to scale to zero... |
Below that, we use CockroachDB dedicated, as like their managed service offering of Cockroach. In the past, we actually did run Cockroach on our own, but we figured the amount of money and time we need to invest basically is being taken by Docker if they run it; it's even economically more value, because they include t... |
And upfront, we use Google CDN and Google Global Load Balancer with Cloud Armor to mitigate some of the risks around DDoS, and rate limits, and stuff, because we see malicious behaviors all the time... And yeah, the stack is really lean in that regard... Because the only thing we can append to that from an operational ... |
**Gerhard Lazu:** Interesting. |
**Florian Forster:** ...because we try to reduce our amount of third-party processes whenever possible on that end. |
**Gerhard Lazu:** Okay. |
**Florian Forster:** That's kind of the stack we use. |
**Gerhard Lazu:** I like that. It's simple, but it sounds modern, and it's complicated in the areas where it's sometimes by design; as you mentioned, this coupling between CockroachDB the dedicated one, which is managed for you, so you're consuming it as a service, and there's a coupling to Datadog. And I'm sure that c... |
**Florian Forster:** Yes... \[laughs\] |
**Gerhard Lazu:** Okay. So there's Google Cloud, there's CockroachDB, there's Datadog. Anything else? |
**Florian Forster:** \[18:10\] We use TerraForm to provision infrastructure, and we use GitHub Actions mainly to do that. We have still some stuff in TerraForm Cloud, but we're constantly migrating into GitHub Actions and private repositories, because it fits better with our flow. 80% of the company knows how to deal w... |
**Break:** \[19:03\] to \[21:02\] |
**Gerhard Lazu:** How much of this do you use for the actual Citadel software as a service, where it's all managed for you? I think you're calling it the cloud now? |
**Florian Forster:** Yup. |
**Gerhard Lazu:** \[unintelligible 00:21:11.10\] versus building the binaries that you make available to users? Is it like most of this is for the cloud service, or what does the combination look like? |
**Florian Forster:** I think 80% is for the cloud service, because the whole CI side of the story comes down to GitHub Actions, some Docker files, and some custom-built shell scripts, because protoc and gRPC tends to be a little bit nifty how to build clients with gRPC. So that basically comes down to that from a CI pe... |
**Gerhard Lazu:** Yeah, of course. |
**Florian Forster:** \[22:23\] Yeah, and you need to reduce the drag. I think an honorable mention on that end is like Cypress for each retest stuff... Because we test our management GUI with Cypress, which basically passes on to our API, so we can do that end to end from our perspective... Yeah, and then there's like ... |
**Gerhard Lazu:** Of course. Of course, that just makes sense. Yeah, that cognitive overload of having ten ways of doing the same thing, based on the part of the company that you're working in... Okay, okay. So how many engineers are there for us to better understand the scale of people that touch this code, and work o... |
**Florian Forster:** Yeah. Currently, it's like eight dedicated engineers, like software engineers from our end, working on Citadel's code. Some of them also working on things like the Helm chart, and stuff, because it's mainly our engineering staff who does that... And then we have like 20, 25-ish external contributor... |
So it's multiple projects, multiple contributors, but the stack beneath is kind of the same. And with us it's basically the eight engineers who work on that, even though the company now has roughly 15 employees since we started to work more and more on technical writing things, API documentation... Because otherwise, i... |
**Gerhard Lazu:** Oh yes, the docs. Oh, yeah. Tell me about it. Yeah. Okay, okay. \[laughter\] Oh, yeah. |
**Florian Forster:** Docs is like the conflicting major pain point all the times. And I don't say that lightly, but I think that you can use that on any project - as soon as it's kind of open source, it's hard to get good quality, consistent reading docs... But that's a topic - for example, we want to invest quite heav... |
**Gerhard Lazu:** Yeah, for sure. And I can't think of a better way to show the users that you care, you care about your product. I mean, very few would dig into the code and say, "Wow, this is amazing code. I'm going to use this product, because it's amazing." How often does it happen? |
**Florian Forster:** \[laughs\] |
**Gerhard Lazu:** But put like a nice docs site out there, easy to search, easy to understand, with good flows, and people go "Wow, they really spent some time", even though the code may not be that good, but the docs are, and then the perception is "This is amazing." |
**Florian Forster:** \[26:02\] Yeah. I mean, I have recently heard the term "Content is oxygen for the community", and I think that's also applicable to the documentation side of things, because it's not only outreach content, but also, and rather important, the documentation side of things. Because even if you write l... |
**Gerhard Lazu:** Yeah. Okay, that makes a lot of sense. Yeah, for sure. For sure. I'm just taken aback by how simply you've put that. It's a very high problem, right? And it all boils down to this. So if you don't get this, then forget everything else, about like making them easy to understand. I mean, that's importan... |
**Florian Forster:** Examples, oftentimes... |
**Gerhard Lazu:** Examples, exactly. |
**Florian Forster:** API documentation... |
**Gerhard Lazu:** Oh, yeah. |
**Florian Forster:** In our case, we split out the whole self-hosting part, because in our cloud it's not applicable. But if you want to run it on your own, you need like production checklists, examples to deploy it to x, y, z, how to configure all the nifty details, how to configure a CDN, how to configure TLS... I me... |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.