build cnpg with AGE and timescale extensions #646

New issue

Closed

opened 2026-03-28 04:27:00 +00:00 by mfreeman451 · 2 comments

mfreeman451 commented

2026-03-28 04:27:00 +00:00

Owner

Imported from GitHub.

Original GitHub issue: #1942
Original author: @mfreeman451
Original URL: https://github.com/carverauto/serviceradar/issues/1942
Original created: 2025-11-14T18:24:27Z

No description provided.

Imported from GitHub. Original GitHub issue: #1942 Original author: @mfreeman451 Original URL: https://github.com/carverauto/serviceradar/issues/1942 Original created: 2025-11-14T18:24:27Z --- _No description provided._

mfreeman451 commented

2026-03-28 04:27:00 +00:00

Author

Owner

Imported GitHub comment.

Original author: @mfreeman451
Original URL: https://github.com/carverauto/serviceradar/issues/1942#issuecomment-3535438129
Original created: 2025-11-15T02:36:20Z

CNPG work is live now:

Pushed the refreshed image with both extension layers (digest ) and added a static tag so we can pin it everywhere.
Updated the Bazel push target, Helm chart, and demo kustomization to consume that tag/digest, plus wired in for GHCR.
Renamed the demo cluster to , applied the manifest, and confirmed are all Ready with both + present in .
Restarted the SPIRE server so it now talks to ; controller manager is reconciling again.
Docs/specs (, agents guide, spiffe identity, OpenSpec) now describe the rename and rebuild flow so we don’t regress.

Next: keep an eye on the live pods for any Timescale/AGE complaints and audit the rest of the demo workloads for lingering hostnames, but the shared CNPG stack is up and running.

Imported GitHub comment. Original author: @mfreeman451 Original URL: https://github.com/carverauto/serviceradar/issues/1942#issuecomment-3535438129 Original created: 2025-11-15T02:36:20Z --- CNPG work is live now: - Pushed the refreshed image with both extension layers (digest ) and added a static tag so we can pin it everywhere. - Updated the Bazel push target, Helm chart, and demo kustomization to consume that tag/digest, plus wired in for GHCR. - Renamed the demo cluster to , applied the manifest, and confirmed are all Ready with both + present in . - Restarted the SPIRE server so it now talks to ; controller manager is reconciling again. - Docs/specs (, agents guide, spiffe identity, OpenSpec) now describe the rename and rebuild flow so we don’t regress. Next: keep an eye on the live pods for any Timescale/AGE complaints and audit the rest of the demo workloads for lingering hostnames, but the shared CNPG stack is up and running.

mfreeman451 commented

2026-03-28 04:27:00 +00:00

Author

Owner

Imported GitHub comment.

Original author: @mfreeman451
Original URL: https://github.com/carverauto/serviceradar/issues/1942#issuecomment-3535438249
Original created: 2025-11-15T02:36:31Z

CNPG work is live now:

Pushed the refreshed image with both extension layers (digest sha256:68f2c4e55b553e7f5f710b28c9c19a56dc37de467543e9b31d4afe002df9d2ac) and added a static 16.6.0-sr1 tag so we can pin it everywhere.
Updated the Bazel push target, Helm chart, and demo kustomization to consume that tag/digest, plus wired in imagePullSecrets for GHCR.
Renamed the demo cluster to cnpg, applied the manifest, and confirmed cnpg-[1-3] are all Ready with both age + timescaledb present in pg_extension.
Restarted the SPIRE server so it now talks to cnpg-rw.demo.svc.cluster.local; controller manager is reconciling again.
Docs/specs (cnpg_01.md, agents guide, spiffe identity, OpenSpec) now describe the rename and rebuild flow so we don’t regress.

Next: keep an eye on the live pods for any Timescale/AGE complaints and audit the rest of the demo workloads for lingering spire-pg hostnames, but the shared CNPG stack is up and running.

Imported GitHub comment. Original author: @mfreeman451 Original URL: https://github.com/carverauto/serviceradar/issues/1942#issuecomment-3535438249 Original created: 2025-11-15T02:36:31Z --- CNPG work is live now: - Pushed the refreshed image with both extension layers (digest sha256:68f2c4e55b553e7f5f710b28c9c19a56dc37de467543e9b31d4afe002df9d2ac) and added a static 16.6.0-sr1 tag so we can pin it everywhere. - Updated the Bazel push target, Helm chart, and demo kustomization to consume that tag/digest, plus wired in imagePullSecrets for GHCR. - Renamed the demo cluster to cnpg, applied the manifest, and confirmed cnpg-[1-3] are all Ready with both age + timescaledb present in pg_extension. - Restarted the SPIRE server so it now talks to cnpg-rw.demo.svc.cluster.local; controller manager is reconciling again. - Docs/specs (cnpg_01.md, agents guide, spiffe identity, OpenSpec) now describe the rename and rebuild flow so we don’t regress. Next: keep an eye on the live pods for any Timescale/AGE complaints and audit the rest of the demo workloads for lingering spire-pg hostnames, but the shared CNPG stack is up and running.