Skip to content

feat: vendor-neutral metrics emission for orchestrate and promote runs #3

@joshua-temple

Description

@joshua-temple

Problem

Workflow runs themselves are the only observability surface. No metrics for orchestrate duration, failure rate, time-in-env, breaking-change gate frequency, or rollback rate. Adopters can't set SLOs on the pipeline itself.

Proposed

Add a metrics emission step inside the generated orchestrate.yaml and promote.yaml finalize jobs, behind a vendor-neutral seam. Output via:

  • A pluggable metrics adapter (e.g., DogStatsD becomes one optional adapter, matching the vendor-neutral telemetry stance) when enabled in the manifest
  • Generic webhook POST (JSON payload) for other backends
  • Always: write metrics as a job summary table for the run UI

Metric set: orchestrate duration, orchestrate outcome (success/failure/skipped), promote duration (from_env, to_env), promote breaking-change gate outcome (allow/block), time-in-env per env.

Impact

Makes the framework observable enough for SLO ownership. Cheap once the finalize step already runs on every cascade.

Schema-reserve note

Design the metrics: field shape (and the adapter seam) into the v1 schema so it's reserved during any freeze. Any specific vendor adapter is one optional implementation behind the seam — no vendor is baked into the core. Implementation can land in a later minor.

Metadata

Metadata

Assignees

No one assigned

    Type

    No type
    No fields configured for issues without a type.

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions