Operations Shared (all languages) active http-servicebatch-job

Per-unit production telemetry is structured and complete

ops-017

Intent

Make requests and jobs diagnosable by emitting one machine-parseable record or root span per unit of work with duration, outcome, and stable identifiers.

Applicability

Applies when the diff adds or changes production request handlers, workers, queue consumers, or cron jobs that emit telemetry.

What to inspect

Entrypoints, structured log payloads, root spans, duration fields, outcome fields, and request or job identifiers.

Pass criteria

Each changed production work unit emits structured completion telemetry with duration, outcome, and a stable correlation identifier.

Fail criteria

The diff emits only narrative log lines or aggregate counters for a work unit, or completion telemetry omits duration, outcome, or stable identifiers.

Do not flag

Low-level helpers already inside an instrumented parent unit of work.

Confidence guidance

HIGH when the entrypoint and incomplete telemetry are directly visible. MEDIUM when a shared wrapper may add fields. LOW when framework auto-instrumentation is unclear.

Remediation

Emit one structured event or root span per work unit and include duration, outcome, and a stable identifier.

Pass example

{
  "request_id": "9f7c1d",
  "path": "/checkout",
  "duration_ms": 184,
  "status": 500
}

Fail example

accepted checkout request
finished checkout request