ADR-023: Public Operational Metrics¶
- Status: Accepted
- Date: 2026-01-11
- Authors: Podcast Scraper Team
- Related RFCs: RFC-026
Context & Problem Statement¶
Understanding the performance of the pipeline (WER, latency, cost, coverage) shouldn't require digging through GitHub Actions logs or running local analysis scripts.
Decision¶
We adopt Public Operational Metrics:
- CI jobs emit a standardized
metrics.jsonartifact. - A dashboard generator converts these into a static HTML site hosted on GitHub Pages.
- Metrics include: Test health, code coverage, and AI pipeline performance.
Rationale¶
- Transparency: Provides an "at-a-glance" view of the project's health for all contributors.
- Trend Analysis: Static hosting allows us to track metrics over months without expensive database infrastructure.
- Accessibility: No special tools or environment required to view the "Scoreboard."
Alternatives Considered¶
- Private Database (InfluxDB/Prometheus): Rejected as overkill and too complex for a single-dev/small-team project.
Consequences¶
- Positive: High visibility; automated progress tracking for the "AI Quality Platform."
- Negative: Requires careful sanitization to ensure no private data (RSS URLs, filenames) leaks into public dashboards.