grafana-util change check
What this page covers
Check whether the staged change package looks structurally safe to continue.
When to open this page
- Run this after change inspect when you need one readiness gate before preview.
- Use this in CI when you want a fast staged readiness decision without building a full plan yet.
- Prefer this over status staged when you want to stay in the task-first change lane.
Who this page is for
Best for review-first workflows where you want summary, preflight, plan, and apply to stay explicit.
Purpose
Check whether the staged change package looks structurally safe to continue.
When to use
- Run this after
change inspectwhen you need one readiness gate before preview. - Use this in CI when you want a fast staged readiness decision without building a full plan yet.
- Prefer this over
status stagedwhen you want to stay in the task-firstchangelane.
Before / After
- Before: you know what the package contains, but not whether the staged inputs are coherent enough to continue.
- After: you have a readiness-style result with blockers and warnings that can stop the workflow early.
Key flags
--workspace: auto-discover the staged package from common repo-local inputs.--availability-file: merge staged availability hints into the check.--target-inventory,--mapping-file: deepen bundle or promotion-oriented checks when those artifacts exist.--fetch-live: merge live target checks into the staged readiness decision.--output-format: render astextorjson.
Examples
# Purpose: Check the discovered staged package.
grafana-util change check --workspace . --output-format json
Expected Output:
{
"status": "ready",
"blockers": [],
"warnings": []
}
# Purpose: Check the staged package with live and availability context.
grafana-util change check --workspace . --fetch-live --availability-file ./availability.json
Expected Output:
PREFLIGHT CHECK:
- dashboards: valid
- datasources: valid
- result: 0 blockers
What success looks like
- the result clearly distinguishes hard blockers from softer warnings
- another operator or CI job can stop safely without reverse-engineering the inputs
- live-backed checks line up with the target environment you meant to inspect
Failure checks
- if blockers appear unexpectedly, verify that staged files and availability hints come from the same environment
- if a live-backed check looks wrong, re-check credentials, org scope, and target URL before trusting the result
- if automation reads the JSON, inspect
status,blockers, andwarningsonly after validating the result shape