TRAILHEAD post-run review¶

Every TRAILHEAD cycle ends with a post-run review. The review determines whether the cycle result is a go (release or promotion can proceed) or a no-go (blockers must be resolved first).

Review output¶

The review produces one of three verdicts:

Review steps¶

1. Read the run log¶

Open tests/trailhead/logs/<date>-<scenario>/run.log. Check:

• All [PHASE] sections completed
• [FAIL] count is zero (or all failures are documented as known/accepted)
• [SKIP] entries have a logged reason
• Close-THRun footer is present (confirms the cycle completed, not aborted)

2. Compare against the prior cycle¶

If a prior cycle exists for the same scenario:

# Compare PASS/FAIL counts between runs
$current = Select-String -Path .\logs\2026-04-13-pre-release-v0.3.0\run.log -Pattern '^\d+:\d+:\d+ \[FAIL\]'
$prior   = Select-String -Path .\logs\2026-03-15-pre-release-v0.2.0\run.log -Pattern '^\d+:\d+:\d+ \[FAIL\]'
"Current failures : $($current.Count)  |  Prior failures: $($prior.Count)"

New failures compared to the prior cycle are regressions and block the release unless explicitly accepted with a written rationale.

3. Verify cluster recovery¶

Confirm the cluster is in the same or better state than before the cycle:

Get-ClusterNode  | Select-Object Name, State
Get-StoragePool  | Select-Object FriendlyName, HealthStatus, OperationalStatus
Get-VirtualDisk  | Select-Object FriendlyName, HealthStatus, ResiliencySettingName

4. Document the verdict¶

Record the verdict in the tracking issue opened by Start-TrailheadRun.ps1. Include:

• Verdict (Go / Go with notes / No-go)
• Link to evidence directory
• Summary of any failures and their disposition
• Reviewer name and date

5. Gate the release (if applicable)¶

If this cycle is a pre-release gate:

• Go: merge the release-please PR and tag the release
• No-go: close the release-please PR, resolve blockers, re-run the cycle