Summary
analyze_kamaji counts pods from a hardcoded kamaji-system namespace (backend/services/scanner/fetch.py:895) and gates DETECTED on running pods — the same class as #202 and the cert-manager sibling. A Kamaji control-plane installed into a non-kamaji-system namespace (e.g. any Helm --namespace) reads 0 pods and mis-reports its status.
Severity: low
Lower likelihood than the cert-manager instance: kamaji-system is the conventional install namespace and there is no widely-shipped platform add-on that relocates it (unlike ROKS/Red Hat cert-manager). Filed so the class is fully tracked, not just the two live instances.
The class
There are exactly three hardcoded-namespace pod fetches:
fetch.py:833 — cert-manager (higher severity, filed separately — downgrades status)fetch.py:834 — kube-systemfetch.py:895 — kamaji-system (this issue)
Suggested fix
Collapse all three into the existing backend/services/bnk_pod_discovery.py:272 _sweep_all_namespaces / list_pod_for_all_namespaces pattern, partitioned by namespace/name. That closes the class permanently and removes the per-component hardcoding.
Context
Follow-up to PR #203 (issue #202, Multus openshift-multus). Raised by @bonnyr-f5 in round-1 review as Major M-1: the multus fix is 1-of-3 of this class; kamaji is the lower-likelihood third instance, kept out of #203's scope.
https://claude.ai/code/session_01UCsZXDxBsWV2s4kT47DwDW
Summary
analyze_kamajicounts pods from a hardcodedkamaji-systemnamespace (backend/services/scanner/fetch.py:895) and gatesDETECTEDon running pods — the same class as #202 and the cert-manager sibling. A Kamaji control-plane installed into a non-kamaji-systemnamespace (e.g. any Helm--namespace) reads 0 pods and mis-reports its status.Severity: low
Lower likelihood than the cert-manager instance:
kamaji-systemis the conventional install namespace and there is no widely-shipped platform add-on that relocates it (unlike ROKS/Red Hat cert-manager). Filed so the class is fully tracked, not just the two live instances.The class
There are exactly three hardcoded-namespace pod fetches:
fetch.py:833—cert-manager(higher severity, filed separately — downgrades status)fetch.py:834—kube-systemfetch.py:895—kamaji-system(this issue)Suggested fix
Collapse all three into the existing
backend/services/bnk_pod_discovery.py:272_sweep_all_namespaces/list_pod_for_all_namespacespattern, partitioned by namespace/name. That closes the class permanently and removes the per-component hardcoding.Context
Follow-up to PR #203 (issue #202, Multus
openshift-multus). Raised by @bonnyr-f5 in round-1 review as Major M-1: the multus fix is 1-of-3 of this class; kamaji is the lower-likelihood third instance, kept out of #203's scope.https://claude.ai/code/session_01UCsZXDxBsWV2s4kT47DwDW