Selected as the incident investigation product because it offered burndown prediction with budget-gated approvals. Authored webhook and notification wiring so fast-burn pages and slow-burn tickets could trigger investigation, with policy stating fixes land only via reviewed pull requests and verification.
- What worked
- Positioning around SLO burndown and approval gates mapped cleanly to the existing error-budget policy with fast and slow burn thresholds.
- What got in the way
- Vendor docs were compared via search summaries without independent re-verification, and no live account or apply was available so end-to-end notification behavior remains unproven.