21–30 of 107 posts
recently facing the dilemma of choosing between k8s vs something more basic. Features that seemed to be advocating for k8s were not server provisionning, but instead : log management, easy setup of blue/green & canary deployment, not having to restart a vm upon new code deployment, etc... How would you do those things as easily with other techs ?
Every other project as different constraints.
> StackOverflow runs on a couple of servers. Does it?
Yes it does: https://stackexchange.com/performance Pretty impressive I think.
That’s not an architecture diagram though, so it doesn’t represent the complexity at all. I’m sure a troubleshooting map for bare linux server wouldn’t be less complicated than that.
When I first saw it K8S reminded me a lot of systemd. I wouldn't be surprised if over the next few years each grow the features of the other.
I hate when stack overflow is held up as an example of how we can run any system on “a few servers” - stacknoverflow has like 3 features and has an engineering focus on the single goal of performance and keeping on running on the small subset of servers. Every other project as different constraints.
I hate when stack overflow is held up as an example of how we can run any system on “a few servers” - stacknoverflow has like 3 features and has an engineering focus on the single goal of performance and keeping on running on the small subset of servers. Every other project as different constraints.
Nowadays, everybody insists on putting stuff on K8s regardless of how large or small it is.
An application is an application for the purpose of running it on a server. It doesn't really matter how much functionality it has.
It is microservices "revolution" (quotes intentionally) that caused larger applications to be split a lot of small ones and complicated the execution environment to the point that a lot of people spend a lot of time just trying to figure out how to run their applications reliably.
That is not necessary.
If you can have multiple microservices, more likely than not you can have them as separate libraries used by single application or a separate modules of a single application. Just make sure to put the same thought into modularizing it as you would designing microservices APIs and you can have the same but much easier and with much better performance (no serialization/deserialization, no network hops, no HTTP stack, no auth, etc.)
Earlier quoted context omitted.
Yes it does: https://stackexchange.com/performance Pretty impressive I think.
No it doesn’t. From your link: • 9 web servers • 4 SQL servers • 2 Redis servers • 3 tag engine servers • 3 Elasticsearch servers • 2 HAProxy servers That comes to 23. I know “a couple” is sometimes used to mean more than two, but… not that much more than two. “A couple” is just flat-out wrong; I’d guess that he’s misinterpreting ancient figures, taking the figures from no later than about 2013 about how many web ser…
* Maybe a bit more/less, because it's not clear to me if DB RAM is per server, or per cluster. Likely server, as on other servers. There is also no data on how big is their haproxy.
I hate when stack overflow is held up as an example of how we can run any system on “a few servers” - stacknoverflow has like 3 features and has an engineering focus on the single goal of performance and keeping on running on the small subset of servers. Every other project as different constraints.
I don't think this really does justice to what Stack Overflow does. They're probably the most visited engineering-related site on any given day. They have their widely known Q&A, live updates and notifications, chat rooms, a blog, job listings with email updates, review queues, moderator tools, and so forth. Perhaps you only use three features, but the site does a tremendous amount, and that's not even thinking about…
Earlier quoted context omitted.
No it doesn’t. From your link: • 9 web servers • 4 SQL servers • 2 Redis servers • 3 tag engine servers • 3 Elasticsearch servers • 2 HAProxy servers That comes to 23. I know “a couple” is sometimes used to mean more than two, but… not that much more than two. “A couple” is just flat-out wrong; I’d guess that he’s misinterpreting ancient figures, taking the figures from no later than about 2013 about how many web ser…
Exactly. That twitter thread is just pure rage based on no data. Sum up resources from that page - we are talking around 6500GB* of RAM worth of servers. That is no homelab. * Maybe a bit more/less, because it's not clear to me if DB RAM is per server, or per cluster. Likely server, as on other servers. There is also no data on how big is their haproxy.
recently facing the dilemma of choosing between k8s vs something more basic. Features that seemed to be advocating for k8s were not server provisionning, but instead : log management, easy setup of blue/green & canary deployment, not having to restart a vm upon new code deployment, etc... How would you do those things as easily with other techs ?
why don't you consider nomad[0] in your evaluations. i think it should fit your requirement. [0] - https://www.nomadproject.io
You'll still get layer-upon-layer of abstraction - for example Consul for key-value and service discovery, Traefik for load balancing, Terraform to build up the service discovery rules, etc - but it feels somewhat more intentful, less boilerplate.