AutoCoder-Engine was deployed with detached Kubernetes Codex worker support and startup issue resume parallelism.
6fc63cc to deploy the detached worker runtime and parallel startup issue resume.1469a1c to enable Kubernetes worker reconciliation via AUTOCODER_K8S_READER_ENABLED = "true" in the repo-managed root config.351d926 to pin worker pods to the deployed backend SHA image, add the registry pull secret, reduce worker resource limits to fit namespace quota, and keep the worker reconciler retrying after per-pod failures.bd21eb9 to allow app=autocoder-agent-worker pods to reach PostgreSQL via NetworkPolicy.k8s/05-deployment-service-ingress.yaml.38e0b9d so future deploys reconcile k8s/04-network-policies.yaml and CI has networkpolicies permissions.562, 564, 565, 566, and 567 completed successfully for the relevant deploy stages.autocoder-engine rolled out image gitea.app.lan/autocoder/autocoder-engine:sha-351d92688a3c.app=autocoder-agent-worker pods were running and heartbeating as Kubernetes workers in codex_worker_leases.The earlier worker failures were caused by latest image pull attempts and Postgres NetworkPolicy restrictions. Both are now managed in repository files.