kubeflow / kubeflow/docs-agent

bug(kagent): Default Kagent bundled agents crash loop due to missing secrets

Open
#89 1 comment 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
Python
Stars
42
Forks
111
Avg merge
6d 23m
Merged PRs (30d)
2

Description

Problem

When installing Kagent without the --set agents.*.enabled=false flags documented in kagent-feast-mcp/README.md (lines 138-149), the default bundled agents (argo-rollouts-conversion-agent, cilium-debug-agent, k8s-agent, etc.) crash loop with CreateContainerConfigError because they expect a kagent-openai secret that doesn't exist in this project's setup.

These agents collectively consume ~2.3GB of RAM before crashing, which on constrained environments (e.g., 8GB WSL2 hosts) prevents the actual kubeflow-docs-agent pod from being scheduled.

Current Mitigation

The README Step 6 already includes the correct --set flags to disable these agents. However:

  1. If a contributor installs kagent from the upstream kagent docs (without these flags), they hit this crash with no explanation.
  2. There's no troubleshooting entry for CreateContainerConfigError or kagent-openai secret in the README.

Proposed Fix

Add a Troubleshooting entry to kagent-feast-mcp/README.md under the existing Troubleshooting section:

### Default agents crash with CreateContainerConfigError

If you installed kagent without the `--set agents.*.enabled=false` flags from Step 6,
the bundled agents will crash because they require a `kagent-openai` secret.
Either reinstall with the flags above, or delete the extra agents:

kubectl delete agent argo-rollouts-conversion-agent cilium-debug-agent \
  cilium-manager-agent cilium-policy-agent helm-agent istio-agent \
  k8s-agent kgateway-agent observability-agent promql-agent -n <YOUR_NAMESPACE>

File

  • kagent-feast-mcp/README.md — Troubleshooting section

Context

Discovered during local Kagent PoC validation — #59 (comment)

Note: @SanthoshToorpu noted that kagent install should remove these.

Confirmed that the --set flags do work — this issue is about documenting the fallback for when those flags are missed.

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Edit kagent-feast-mcp/README.md in the existing Troubleshooting section. First compare the proposed entry with the Step 6 installation flags and confirm the listed agents and namespace placeholder; done means the README explains the CreateContainerConfigError and kagent-openai secret problem and documents both recovery options.

Written by the indexing model from the issue text.

Assessment

Tech stack
kubernetes
Domain
documentation
Issue type
Documentation
Difficulty
1/5
Estimated time
Under an hour
Activity status
Stale
Clarity
Clearly specified
Newbie friendliness
50/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.