chore(ci): wait for Grafana Cloud readiness before dashboard provisioning

[FSM1]

Summary

• Grafana Cloud free-tier instances sleep when idle and return 503 "Your instance is loading" on first access
• Adds a readiness poll step (up to 5 min, 10s intervals) against /api/health before pushing the dashboard
• Fixes consistently failing Provision Grafana Dashboard job in deploy-staging workflow

Test plan

• Trigger a staging deploy and verify the readiness step waits then succeeds
• Verify dashboard is provisioned after the wait completes

🤖 Generated with Claude Code

Summary by CodeRabbit

• Chores
  • Improved deployment workflow: added a dashboard provisioning step that normalizes the target URL, waits for Grafana to become healthy with retries, and fails fast on authentication/permission or not-found errors to prevent faulty pushes.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: 314780e2-9f53-45b8-bc62-856df5e06303

📥 Commits

Reviewing files that changed from the base of the PR and between 440cd51 and fa6ad3b.

📒 Files selected for processing (1)

• .github/workflows/deploy-staging.yml

🚧 Files skipped from review as they are similar to previous changes (1)

• .github/workflows/deploy-staging.yml

---

Walkthrough

Added a GitHub Actions workflow step (provision-dashboard) that normalizes GRAFANA_URL and polls GET {GRAFANA_URL}/api/health with a Bearer GRAFANA_API_KEY, retrying up to 30 times at 10s intervals; succeeds on HTTP 200, fails immediately on 401/403/404, otherwise continues polling.

Changes

Cohort / File(s)	Summary
Grafana Cloud readiness & provisioning .github/workflows/deploy-staging.yml	Inserted a provision-dashboard step before "Push dashboard to Grafana Cloud": normalizes GRAFANA_URL, performs health polling against GET /api/health with Authorization: Bearer $GRAFANA_API_KEY, retries up to 30 times with 10s sleeps, treats 200 as success, 401/403/404 as immediate failures, and errors out after timeout.

Estimated code review effort

🎯 2 (Simple) | ⏱️ ~10 minutes

Possibly related PRs

• fix: auto-provision Grafana dashboard on staging deploy #282: Adds the provision-dashboard job that pushes Grafana dashboards; this change inserts a readiness check before that push.
• docs: add Grafana Cloud dashboard provisioning learnings #287: Documents Grafana provisioning edge cases (trailing-slash normalization, 403 behavior) and relates to the URL normalization and status handling added here.

🚥 Pre-merge checks | ✅ 3

✅ Passed checks (3 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title accurately describes the main change: adding a readiness check for Grafana Cloud before dashboard provisioning in the CI workflow.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch chore/ci-grafana-readiness-wait

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[Copilot]

Pull request overview

This PR improves the reliability of the staging deployment workflow by adding a Grafana Cloud readiness check to handle free-tier “sleep/warm-up” behavior before attempting dashboard provisioning.

Changes:

• Adds a polling step (up to 5 minutes) that waits for Grafana Cloud to return HTTP 200 on /api/health.
• Ensures the “Push dashboard to Grafana Cloud” step runs only after Grafana is responsive.

[coderabbitai]

Actionable comments posted: 1

🧹 Nitpick comments (1)

.github/workflows/deploy-staging.yml (1)

514-516: Fail fast on missing Grafana env vars.

This step should validate required env vars up front; otherwise it retries for minutes on a configuration error.

💡 Proposed fix (guard clauses)

           set -euo pipefail
+          : "${GRAFANA_URL:?GRAFANA_URL is required}"
+          : "${GRAFANA_API_KEY:?GRAFANA_API_KEY is required}"
           GRAFANA_URL="${GRAFANA_URL%/}"

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In @.github/workflows/deploy-staging.yml around lines 514 - 516, Add upfront
guard clauses after "set -euo pipefail" to validate required Grafana environment
variables (e.g. GRAFANA_URL and GRAFANA_API_KEY) before entering retry logic;
check that GRAFANA_URL is non-empty (after the existing
GRAFANA_URL="${GRAFANA_URL%/}" normalization) and that GRAFANA_API_KEY (or any
other required var) is set, and if any are missing emit a clear error to stderr
and exit non-zero immediately (avoiding long MAX_ATTEMPTS retries). Ensure
messages reference the variable names so the failure is actionable.

🤖 Prompt for all review comments with AI agents

Verify each finding against the current code and only fix it if needed.

Inline comments:
In @.github/workflows/deploy-staging.yml:
- Around line 518-531: The retry loop uses a fixed MAX_ATTEMPTS which can exceed
a 5-minute wall-clock limit because each curl has its own timeout and there's
sleep; change the loop to enforce a deadline-based timeout instead: capture a
start time (or compute a deadline = now + 300), replace the for ((i=1...)) loop
with a loop that checks elapsed time against 300 seconds (or compares now to
deadline) and increments attempt counter only for logging, run the curl + sleep
until either HTTP 200 or elapsed >= 300, and update the final error/log to
report that Grafana did not become ready after the 5-minute deadline (use the
same variables like HTTP_CODE, INTERVAL, MAX_ATTEMPTS only for logging as
needed).

---

Nitpick comments:
In @.github/workflows/deploy-staging.yml:
- Around line 514-516: Add upfront guard clauses after "set -euo pipefail" to
validate required Grafana environment variables (e.g. GRAFANA_URL and
GRAFANA_API_KEY) before entering retry logic; check that GRAFANA_URL is
non-empty (after the existing GRAFANA_URL="${GRAFANA_URL%/}" normalization) and
that GRAFANA_API_KEY (or any other required var) is set, and if any are missing
emit a clear error to stderr and exit non-zero immediately (avoiding long
MAX_ATTEMPTS retries). Ensure messages reference the variable names so the
failure is actionable.

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: 6fd139d9-1814-4244-a2c0-53f8fd13ae33

📥 Commits

Reviewing files that changed from the base of the PR and between e35f6ba and 440cd51.

📒 Files selected for processing (1)

• .github/workflows/deploy-staging.yml