Clarification on update strategy

[JakubSzuber]

Hi, I'm opening this issue because I want to clarify something, and this repo doesn't have discussions enabled.

According to the docs the default update startegy is rollout, and the docs warns that this strategy cannot be used with GitOps agent, as Reloader will patch the deployment, and e.g. Argo CD will see some drift => try to sync.

I test this out on my setup with Argo CD (auto-sync enabled) and the default Reloader startegy (rollout), and everything seems to work perfectly fine!

On ConfigMap update, the Reloader updates (patch) value of STAKATER_TS_FRONTEND_CONFIGMAP variables in the deployment manifest, so the rolling update is performed and Argo CD does nothing because ArgoCD not reporting diff in environment variable configuration, so there is no sync done after patch from Reloader.

Everything works without restart strategy, so either this section is wrong and requires a fix, or I miss something.

[msafwankarim]

Hi! What is the workload type that you're using? Documentation is pretty ambiguous about this but rollout strategy is only applicable for ArgoCD rollouts. For kubernetes native workload types (Deployments, StatefulSets etc.) we have reload stratigies with possible values of env-vars or annotations

[JakubSzuber]

Hi Muhhamed, I'm using a standard K8s Deployment running on AWS EKS (Deployment is within a Helm chart that is synced by Argo CD). I guess it's not relevant, but I use Argo CD ApplicationSet along with progressive sync.

So it would be cool if you could update the docs here and especially here because it seems like it's absolutely ok to use a default rolling strategy (env-vars) with GitOps agents like Argo CD. Current docs is very confusing.

Again, refer to this GH issue for explanation by it's working fine.