[Fix] Protect TimescaleDB first boot with startupProbe

[tal-haim]

📝 Description

Stops TimescaleDB from being killed by the liveness probe during first-boot initdb on slow storage.

Without a startupProbe, Kubernetes can interrupt initialization and leave a corrupt PGDATA (missing global/pg_filenode.map), causing a permanent crash loop that breaks model-monitoring flows.

This change adds a configurable TimescaleDB startupProbe and forces probe checks over TCP (-h 127.0.0.1).

---

🛠️ Changes Made

• charts/mlrun-ce/templates/timescaledb/statefulset.yaml: added startupProbe; added -h 127.0.0.1 to startup/liveness/readiness probes; removed initialDelaySeconds from liveness/readiness (gated by startup probe)
• charts/mlrun-ce/values.yaml: added timescaledb.startupProbe defaults (periodSeconds: 10, timeoutSeconds: 5, failureThreshold: 30 → ~5 minute startup window)
• charts/mlrun-ce/Chart.yaml: bumped version 0.12.0-rc.4 → 0.12.0-rc.5

---

✅ Checklist

• I have tested the changes in this PR
• I confirmed whether my changes require a change in documentation and if so, I created another PR in MLRun for the relevant documentation.
• I confirmed whether my changes require changes in QA tests, for example: credentials changes, resources naming change and if so, I updated the relevant Jira ticket for QA.
• I increased the Chart version in charts/mlrun-ce/Chart.yaml.
• I confirmed that the installation works both on a local Docker Desktop environment and on a real cluster when using the required prerequisites.
  • If installation issues were found, I updated the relevant Jira ticket with the issue and steps to reproduce, or updated the prerequisites documentation if the issue is related to missing or outdated prerequisites.
• If needed, update https://github.com/mlrun/ce/blob/development/charts/mlrun-ce/README.md with the relevant installation instructions and version Matrix.
• If needed, update the following values files for multi namespace support:
  • Admin values
  • User values Node Port
  • User values ClusterIP

---

🧪 Testing

• helm lint charts/mlrun-ce — passed
• ./tests/helm-template-test.sh — 82/82 passed
• Full cluster make tests / Kind install — not run
• Manual reproduction of mid-initdb kill failure mode was documented on the Jira ticket

---

🔗 References

• Ticket link:
• External links:
• Design docs links (Optional):

---

🚨 Breaking Changes?

• Yes (explain below)
• No

---

🔍️ Additional Notes

Install-mode values files were not updated because timescaledb.startupProbe is a default inherited from values.yaml; admin install keeps timescaledb.enabled: false, and non-admin overlays do not need a probe override.

Jenkins deploy health-gate for TimescaleDB (ticket item 3) still lives in ce-deployment and is out of scope for this PR.

[royischoss]

Hey looks good to me small comment

[royischoss]

LGTM