feat: change docker to use "coder" user and add basic Helm chart #2746
Conversation
| protocol: TCP | ||
| - name: https | ||
| port: 443 | ||
| targetPort: https |
There was a problem hiding this comment.
I don't think we change the deployment port we listen on, so we can't route to 443 here.
There was a problem hiding this comment.
we should probably fix that and provide a second listen address for https traffic. I'll open an issue
helm/templates/ingress.yaml
Outdated
| {{- if .Values.coder.ingress.enable }} | ||
| --- | ||
| apiVersion: networking.k8s.io/v1 | ||
| kind: Ingress |
There was a problem hiding this comment.
Do we need to provide an ingress? Would a service be enough? I think @f0ssel had some thoughts here.
There was a problem hiding this comment.
The Ingress is useful for customers who don't want to expose coder directly with a LoadBalancer Service but instead want to run it through their own ingress controller like the nginx one. Turning on the ingress does not create an ingress controller for you like it did in v1 at one point, the customer needs to supply their own.
There was a problem hiding this comment.
Would it be simpler to exclude that resource for now, as customers can really easily add their own Ingress?
My intent is to minimize the surface area and knowledge of Kubernetes that we need to have. Reducing resources that some devs may not understand can be helpful there.
There was a problem hiding this comment.
Sounds good to me, I'm down to remove it if you're fine with that in the name of simplicity
There was a problem hiding this comment.
I'm good with that if you feel it's not necessary, I don't have a full understanding, I'm just tryna simplify ;p
There was a problem hiding this comment.
@f0ssel you chill with me removing ingress?
There was a problem hiding this comment.
It will be required for us to document the step of optionally creating an ingress on setup instructions since it will be a super common ask, but I'm fine with it not being a part of coder chart. This will also mean "rolling out changes" to an ingress will also be more complicated to instruct users, but I don't see that being something we really need to worry about imo.
| #- name: CODER_PG_CONNECTION_URL | ||
| # value: "postgres://coder:password@postgres:5432/coder?sslmode=disable" |
There was a problem hiding this comment.
What is our recommended setup in this deployment model?
It'd be awesome if you wrote docs as part of this PR (or in a future one, but soon after this). I'm not entirely sure what happens after I helm install this. I believe it just fails due to no PostgreSQL cluster?
There was a problem hiding this comment.
Made an issue about writing documentation: #3224
There was a problem hiding this comment.
at the moment it would just fail unless you specified a postgres deployment URL like that
Adds a basic Helm chart that we can build upon. This basic Helm chart does not contain a postgres database sidecar, and I don't think we should add one as the timescale database in v1 was only for evaluation purposes, but we have better evaluation methods like the built-in postgres for coder v2.
Subtasks
add release publishing of helm chart(I've decided to defer this, right now they get uploaded as a release artifact Publish helm artifacts to a repository #3127)Fixes #2267