Jev example · Engineering & operations

Errors immediately after a deploy

Use a change correlation without assuming proof of cause.

Workflows & decisionschoicenoulscore

State

This is the context Jev receives before answering the typed question.

{
  "report": "Checkout errors increased after release 2.8.",
  "metrics": {
    "error_rate_before": "0.1%",
    "error_rate_now": "18%",
    "duration_minutes": 12
  },
  "scope": "All regions",
  "recent_change": "Release 2.8 deployed 15 minutes ago"
}

Typed questions (3)

Triage route

choice

Route the issue using current impact and supporting evidence.

Try in Playground
  • incident | Active user-facing outage or substantial degradation
  • bug | Reproducible defect without broad active outage
  • planned_work | Known maintenance or expected change
  • investigate | Incomplete or conflicting evidence

Concrete evidence

noul

Does the state provide a specific observation, measurement, or reproduction path that supports the report?

Try in Playground

Service impact

score

Rate the demonstrated present impact.

Try in Playground
  • No service impact
  • Limited inconvenience
  • One important workflow affected
  • Broad service unavailable

Try changing the scenario

Change the errors to a single test account.

Catalog source: TypeSafeAI/typesafe-playground, reused under MIT. This is an independent community catalog, not an official TypeSafe AI product.