Operations and Observability

Operations and Observability

Stub. Tracked as a "still to write" page on the Home hub.

Local-first telemetry, daemon health, and (once cloud is GA) cloud-side operations:

• Local self-checks — budi doctor's three attribution checks (session visibility, branch attribution, activity attribution), legacy-proxy-residue detection, statusline wiring check
• Cost-confidence semantics — what the ~ prefix in dashboards means, when cost_confidence is estimated vs exact vs estimated_unknown_model
• Pricing observability — GET /pricing/status, budi pricing status text output, rejected_upstream_rows[] surface, manifest_version tagging on pricing_source
• Cloud-side ops (post-GA) — Datadog index, OpsGenie recipient, Slack alert channel, SLOs
• Schema audits — budi db check, budi db check --fix, what schema drift looks like and how it's repaired
• Legacy residue reporting — proxy_estimated rows in messages, retained-state reporting in budi doctor

Current source of truth

• SOUL.md — Build & Test, Database, Statusline contract sections
• crates/budi-cli/src/commands/doctor.rs
• crates/budi-daemon/src/routes/pricing.rs
• scripts/e2e/ — end-to-end guards (each script pins a specific contract or bug; scripts/e2e/README.md is the index)