- Ya
- 4 Septemba 2026, 22:13 UTC
- Mwandishi
- Kamo
- Ahadi ya
- e6dabb1
kamo-register is the only service on the platform whose CI generates its Deployment inline, with a heredoc, instead of applying k8s/deployment.yaml. Two consequences, both silent: 1. k8s/deployment.yaml is dead code. The fleet-wide rollout-safety work landed in it this morning like everywhere else, the deploy went green, and nothing reached the cluster. 2. Every deploy REVERTS the live Deployment to that minimal spec, so anything applied by hand lasts until the next push. Caught by the KlusterServices deploy-safety audit, which reads the cluster rather than the repos — this deployment was left as one of only three failures, running with no readiness probe, no preStop hook and no minReadySeconds while its own manifest declared all three. A per-repo guard would have passed here, because the file it would have checked is the file nobody applies. The settings are added to the heredoc, which is the honest place for them today. Switching the step over to k8s/deployment.yaml is NOT a one-line change and is deliberately not attempted here: that file's envFrom references a kamowebregister-config ConfigMap which does not exist in the cluster and which nothing creates, and its env entries are config the running app has never had. Applying it as-is would take the service down. The file now carries a header saying all of this so the next person does not lose the same afternoon to it.