Helm chart update #695
Helm chart update #695
Conversation
EronWright
added
the
impact/no-changelog-required
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## v2 #695 +/- ##
=======================================
Coverage 49.23% 49.23%
=======================================
Files 26 26
Lines 2890 2890
=======================================
Hits 1423 1423
Misses 1290 1290
Partials 177 177 ☔ View full report in Codecov by Sentry. |
| version: 0.9.0 | ||
| appVersion: 1.14.0 | ||
| version: 2.0.0 | ||
| appVersion: "v2.0-devel" |
There was a problem hiding this comment.
Rationale for changing the version info:
appVersionis used as the image tag in deployment.yaml and should match our normal version string. The prerelease target will replace the value with the tag.versionis the chart version, and I would advocate for doing a major bump.
| createClusterRole: false | ||
|
|
||
| ## -- Enable this and set the rules: to whatever custom rules you want for the Cluster Role resource. | ||
| clusterRoleRules: |
There was a problem hiding this comment.
Update: I think I know the original purpose of the ClusterRole, it was to grant admin-level access to the Pulumi program and its Kubernetes Provider, not simply to grant the requisite access to the operator. Now I have repurposed it to support a cluster-wide installation of the operator. The Pulumi program runs in its own service account and user-controlled bindings.
So createClusterRole and clusterRoleRules are obsolete. Under the new rbac section there's some flags related to the operator's rbac.
| # -- Specifies whether cluster roles and bindings should be created | ||
| createClusterRole: true | ||
| # -- Specifies whether namespaced roles and bindings should be created | ||
| createRole: false |
There was a problem hiding this comment.
one either creates a clusterrole or a role for the controller manager, depending on whether you want a namepaced installation of the operator. There's still some follow-up work needed (#690), but the helm piece is in place.
| cpu: 500m | ||
| memory: 5123Mi | ||
| cpu: 200m | ||
| memory: 128Mi | ||
| requests: | ||
| cpu: 100m | ||
| cpu: 200m |
There was a problem hiding this comment.
Rationale for changing the resources is to ensure that the pod is assigned the "Guaranteed" QoS class.
There was a problem hiding this comment.
Awesome!
Obviously not something that we need to resolve now but it would be nice to more easily keep this in sync with the generated configs. There's kubernetes-sigs/kubebuilder#3074 which took me to https://github.com/arttor/helmify which looks interesting.
@blampe the original helm chart for v1 was actually scaffolded using |
Proposed changes
Updates the Helm chart to install PKOv2, as similarly as possible to
operator/config/default.Details:
To install:
Related issues (optional)
Closes #684