Hi, I am recently learning opamp-go and its supervisor. May I know if supervisor is only needed for local collector management, typically, a non-K8s env, such as on a host. In k8s env, what's the recommended solution for collector management?
By reading the doc and also the code in this repo, it looks the supervisor is usually co-located with the collector and it has both direct read and write permission to the collector, e.g.: read and update the configuration as a local file. However, it doesn't mention how that fits a k8s env (or maybe I missed it), where in a k8s env, the collector configuration is usually represented as a ConfigMap or a Custom Resource (managed by the operator).
Also, when it comes to GitOps, e.g.: both the collector deployment and configuration are maintained in a git repo and synced by Argo CD. There's no way to directly touch either the deployment or the configuration in runtime as it will be reverted soon by Argo CD.
With that, I'd like to know if there's any clear recommendation or best practices for this? Thanks!
P.S.: I noticed that there's an OpAMP Bridge implemented in opentelemetry-operator project, which seems very similar to supervisor in a k8s env. Or, if I have the bridge installed, do I still need supervisor?
Hi, I am recently learning opamp-go and its supervisor. May I know if supervisor is only needed for local collector management, typically, a non-K8s env, such as on a host. In k8s env, what's the recommended solution for collector management?
By reading the doc and also the code in this repo, it looks the supervisor is usually co-located with the collector and it has both direct read and write permission to the collector, e.g.: read and update the configuration as a local file. However, it doesn't mention how that fits a k8s env (or maybe I missed it), where in a k8s env, the collector configuration is usually represented as a ConfigMap or a Custom Resource (managed by the operator).
Also, when it comes to GitOps, e.g.: both the collector deployment and configuration are maintained in a git repo and synced by Argo CD. There's no way to directly touch either the deployment or the configuration in runtime as it will be reverted soon by Argo CD.
With that, I'd like to know if there's any clear recommendation or best practices for this? Thanks!
P.S.: I noticed that there's an OpAMP Bridge implemented in opentelemetry-operator project, which seems very similar to supervisor in a k8s env. Or, if I have the bridge installed, do I still need supervisor?