Prevent repeated refreshes from exhausting incident budgets
Several operators refresh the same incident and trigger duplicate model calls.
- Focused work estimate
- 2h 30m + prerequisites
- Priority in the scenario
- High
- Engineering practice
- Cost controls
Estimated field mix
- Applied AI60%
- Backend40%
Field percentages are editorial estimates of the ticket's engineering focus. They total 100%; they are not measured time, proficiency scores, or ownership evidence.
Review it, then add it to your workspace.
The board opens an editable draft; nothing is saved until you confirm it. Sign-in and workspace permissions apply, and Demo boards remain ephemeral.
Project context
A fictional on-call team wants an assistant that joins alerts, deployment notes, and runbooks. Incomplete telemetry and hostile log content make unsupported conclusions dangerous.
Setup prerequisites
- Author synthetic alerts, deployment records, and runbooks plus deterministic tool and model doubles.
Preceding work
Complete these dependencies, or supply their agreed outputs before taking this ticket.
- BTRIAGE-101 · Define a common timeline record for operational events
- BTRIAGE-102 · Limit incident retrieval to the caller's service grants
- BTRIAGE-103 · Bound tool arguments and returned telemetry volume
- BTRIAGE-105 · Require source references for incident summary statements
- BTRIAGE-107 · Cancel obsolete drafts when new incident context arrives
Acceptance criteria
- Coalesce identical in-flight context requests.
- Enforce per-incident reservation and timeout.
- Settle usage once per provider attempt.
Implementation constraints
- Provider failure cannot trigger unbounded retry loops.
Verification to include
- Share one in-flight synthetic response.
- Exhaust the budget and return a bounded unavailable state.
Deliverables
- Incident request budget.
Rollout and recovery
Default to deterministic mode; pause new refreshes when accounting is uncertain.
Value of the work
For the engineer: Practice constrained tool use, operational uncertainty, and traceable AI outputs.
For the team: Produce concise incident drafts that preserve source traceability and operator control.
Evidence boundaries
Outcome Evidence: Tests, patches, and runbooks are requested deliverables. They become Outcome Evidence only through a qualified Mission and immutable Evidence IDs.
Ownership Evidence: Independent adaptation must be observed under a declared verification policy and cite immutable Evidence IDs. Completing a planning ticket establishes no Ownership Evidence.