Operations handbook (weekly health + incident + upgrades)

This handbook defines the recurring operator workflow for running SDETKit as a platform.

1) Weekly health review

Cadence: every Wednesday.

Checklist:

1. Review KPI board from the Top-tier program dashboard.
2. Confirm each workstream has owner, status, and evidence links.
3. Identify top 3 risks and assign mitigation owners.
4. Publish a mid-week status note.

2) Incident triage flow

1. Detect signal (failed release gate, doctor regression, schema drift).
2. Assign severity via Support and escalation model.
3. Start incident record and assign incident commander.
4. Mitigate and communicate using severity update cadence.
5. Close incident with root cause and follow-up tasks.

3) Upgrade planning lane

Run this monthly before release train updates:

1. Review Policy compatibility matrix.
2. Inventory legacy usage via migration tooling/docs.
3. Run canonical path in staging repos.
4. Publish upgrade plan (timeline, owners, rollback plan).
5. Approve rollout wave sequence (startup -> scale -> regulated).

4) Release owner checklist

For every version cut:

• Validate gate fast and gate release artifacts are attached.
• Confirm doctor output is healthy.
• Confirm docs navigation reflects new/changed controls.
• Confirm compatibility/deprecation notes are updated.
• Confirm support/escalation notices are prepared if needed.

5) Evidence policy

Every operational claim should reference one of:

• canonical gate artifacts
• dashboard status entries
• incident records
• release notes with migration guidance