Self-hosted troubleshooting and incident response

Self-hosted troubleshooting and incident response

Self-hosted gives more control and also more ownership. This page is a fast operational playbook: what to do when something fails, and how to reduce repeat incidents by capturing lessons learned.

Purpose and prerequisites

• Purpose: stabilize operations, determine root cause, restore state, and preserve audit trace.
• Prerequisites: access to self-hosted instance, admin/workspace-owner rights, run history access.
• Tools: runbook, previous logs, current template, and target status.

Incident response checklist in steps

1. Immediate stabilization: stop additional changes, narrow scope, confirm rollback option.
2. Collect facts: run ID, workspace, target server, change type, last known healthy state.
3. Check preflight and logs: review preflight result and log stream.
4. Impact assessment: check service availability and security impact.
5. Recovery choice: fix inputs, re-run, or rollback according to playbook.
6. Communication: update team, add context to run comments.
7. Prevention: update checklist and strengthen checks for next run.

Execution checklist

• Have you identified the correct run ID and workspace target?
• Do you have preflight + runtime logs before and after the incident?
• Have you validated template dependencies and variables?
• Is workspace access reviewed (who could execute this change)?
• Can you classify whether this is a one-off issue or repeated drift?

Common issues

• Rerunning without fixing the root cause.
• Retrying while missing dependency or inventory mismatch.
• Incident documentation missing key metadata.
• No cross-team communication during recovery.

Next steps

• First deployment and production checklist before repeat execution.
• Runbook and governance as mandatory reference for risky changes.
• Templates and workflow to improve preflight and operational guardrails.
• Workspaces and access when ownership needs adjustment.

Cross-track links

If you are considering moving to managed mode, continue with mode migration or compare managed vs self-hosted.

Next step: Go to first deployment or the runbook for day-to-day operations.