Introduce a readiness endpoint that tracks metric stabilization

[mraerino]

Description

ref https://linear.app/supabase/issue/NET-510/cloudflare-exporter-delay-new-pod-readiness-until-metrics-stabilize

Adds a readiness HTTP endpoint that can be used in a readiness probe to keep the old pod around until the new pod has stabilized its push metrics

[Copilot]

Pull request overview

Adds an HTTP readiness endpoint intended for Kubernetes readiness probes, so a new pod can be held “not ready” until the VictoriaMetrics push-based (converge) metric pipelines have stabilized.

Changes:

• Introduces readinessTracker and a /ready JSON endpoint that returns 200 only after required components have stabilized.
• Wires converge runners to notify readiness once stabilization occurs.
• Enables an optional Helm readinessProbe (default enabled) that probes GET /ready, and adds a basic HTTP test for the new endpoint.

Reviewed changes

Copilot reviewed 8 out of 8 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
tests/basic_tests.yml	Adds a basic HTTP check for /ready.
readiness.go	Implements readiness tracking and an HTTP handler emitting JSON readiness state.
readiness_test.go	Unit tests for readiness tracking stickiness/idempotency.
main.go	Instantiates readiness tracker, exposes /ready, and passes tracker into converge setup.
converge.go	Adds a stability notifier into converge setup and forwards it to converge.Run.
charts/cloudflare-exporter/values.yaml	Adds configurable readinessProbe values and docs.
charts/cloudflare-exporter/templates/deployment.yaml	Conditionally renders a container readinessProbe for /ready.
charts/cloudflare-exporter/Chart.yaml	Bumps chart version to 0.2.3.

Suppressed comments (1)

main.go:288

• The TODO comment here becomes unnecessary (and the current "todo" prefix may be flagged by comment/style linters). If readiness gating is driven solely by whether vm_push_endpoint is configured, this can be expressed without the TODO block.

	// todo: this is a bad way to check for intended push failure
	// ideally converge setup wouldn't be attempted with this option empty
	if viper.GetString(argVMPushEndpoint) != "" {

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[Copilot]

Pull request overview

Copilot reviewed 9 out of 9 changed files in this pull request and generated 1 comment.