The pattern: monitor Deployment Manager during active deployment windows, use per-component results to debug failures, deploy quick-deploys for tested validations, retain artifacts for compliance. The page is operational furniture; using it actively prevents many deployment surprises.
- Open Deployment Manager during active deployment windows
Setup, Deployment, Deployment Manager. Watch in-progress deployments; catch failures as they happen.
- Drill into failed deployments for per-component errors
Click any failed deployment to see per-component results. The error messages drive fixes.
- Use Quick Deploy for validated deployments during production windows
Validate first, then Quick Deploy within the validation expiration. Skips re-running tests during the production window.
- Cancel runaway deployments deliberately
Cancel is the emergency button. Use only when the deployment is clearly going wrong; expect partial-state cleanup after.
- Correlate with source-tool logs (CI, DevOps Center)
Deployment Manager shows the platform-side view; CI logs and DevOps Center show the originating-tool view. The combination gives the full picture.
- Archive deployment manifests to long-term storage for compliance
Source-tracked deploys produce manifests; change sets produce component lists. Archive both for compliance audit retention.
- Build alerts on failed deployments
Flow on DeployRequest or a CI-side hook that posts to Slack on failure. Manual review catches most failures eventually; alerts catch them immediately.
Change Set, CLI, DevOps Center, Third-Party, Metadata API. Each surfaces in Deployment Manager with consistent metadata.
Validate Only runs the deployment without committing; Deploy commits. Validated deployments support Quick Deploy.
Skips test re-execution on previously-validated deployments. Saves time during production windows.
Emergency abort for in-progress deployments. May leave partial state.
Deployment manifests archived to long-term storage for compliance evidence.
- Deployment Manager retention is limited (days, not months). Long-term audit requires SetupAuditTrail plus tool-side log retention.
- Quick Deploy expiration (typically 10 days from validation) is firm. Validation that lingers past expiration requires re-validation before deploy.
- Cancel can leave partial state. Some changes commit before the cancel completes; manual cleanup follows.
- Per-component error messages are the diagnostic; the deployment-level summary is often too generic. Always drill into per-component for failed deployments.
- Source-tool logs are the source of truth for CI-initiated deployments. Deployment Manager is the platform-side view; both are useful, neither is complete alone.