Health-probe docs recommend `curl -fsS` for renderer readiness, which breaks container startup

[dzirtusss]

Area: react-on-rails-pro-node-renderer — health endpoints (RENDERER_ENABLE_HEALTH_ENDPOINTS=true)

Summary

The recommended Control Plane / Kubernetes readiness probe command for the node-renderer health endpoints is:

curl --http2-prior-knowledge -fsS http://localhost:3800/ready

Configured exactly as documented, the readiness probe fails and the container never becomes ready / cannot start. Removing -fsS makes the probe pass and the container starts normally.

What works

curl --http2-prior-knowledge http://127.0.0.1:3800/info

Root cause (as observed)

The breakage is caused by the -fsS flags, not by the endpoint choice (/ready vs /info). Specifically -f/--fail, which makes curl exit non-zero on any HTTP status ≥ 400 — so whatever non-2xx behavior the renderer's health endpoint exhibits is turned into a failed probe and a container that won't come up.

Request

1. Document the readiness probe command that actually works with the renderer's health endpoints (the docs/recommendation should not include -fsS unless the endpoints are guaranteed to return 2xx in all probe conditions).
2. Clarify the intended status-code contract of /ready (and /info) so probe configs can be set correctly.
3. Note the platform context: on Control Plane there are only two probe methods relevant here — HTTP and Command (exec) probes. The current docs/guidance describe configuration options and HTTP/HTTP2 behavior that don't map to what CPLN actually exposes; please align the docs with the real probe options.

Environment

• RENDERER_ENABLE_HEALTH_ENDPOINTS=true
• Renderer served over h2c (hence --http2-prior-knowledge)
• Platform: Control Plane (CPLN)