Introduce troubleshooting documentation and reduce chance of slow connectivity problems

[pkosiec]

Overview

Acceptance Criteria

Docs

• Docs: Describe how to diagnoze some problems (e.g. get logs, kubectl describe etc.) (something like https://capact.io/docs/operation/diagnostics)

  • how to get logs
  • Use config-exporter to dump the configuration for all bug reports
    • See how it is used in https://github.com/kubeshop/botkube/blob/main/cmd/cli/cmd/migrate.go
    • Guide user to run the job on K8s cluster and get the result

• Docs: Describe common problems (like https://capact.io/docs/operation/common-problems)

  • e.g. Botkube killed by Kubernetes when internet connectivity is too slow
  • Bot not invited to channel - probably it's already fixed by our CLI
  • Donna and Maria had the problems
  • Encourage to join community slack in case someone is stuck and is looking for help

Botkube Open Source

• CLI: Ensure we link to the troubleshooting docs in CLI installation failed message
  • Maybe replace "Installation failed" error message to something different? It's not always about installation, it could be a temporary problem with a pod.
• Agent: Consider increasing thresholds for the probes as plugin download might take a while (e.g. to 50? 100?)
  • https://github.com/kubeshop/botkube/blob/main/helm/botkube/values.yaml#L949
  • https://github.com/kubeshop/botkube/blob/main/helm/botkube/values.yaml#L962
• Goreleaser config for plugins: release plugins in tar.gz archives
  • For local env keep binaries anyway
  • We use go-getter so unpacking that will be out of the box (just need to double check they are packed without additional nested directory)

Reason

Currently we lack a troubleshooting-related guide, which helps with basic debugging. Sometimes on our community Slack we receive similar questions, and we could have a single place with most common issues.