# Reachability operations

Purpose
- Operate call graph ingestion, reachability computation, and explain queries.

Reachability statuses
- unreachable, possibly_reachable, reachable_static, reachable_proven, unknown.

Call graph operations
- Upload call graphs and validate schema.
- Inspect entrypoints and merge graphs when required.
- Enforce size limits and deterministic ordering.

Computation
- Trigger reachability computation per scan or batch.
- Monitor jobs for timeouts and memory caps.
- Persist results with graph_cache_epoch for replay.

Explain queries
- Explain a single finding or batch.
- Provide alternate paths and reasons for unreachable results.

Drift handling
- Track changes due to graph updates or reachability algorithm changes.
- Use drift reports to compare runs and highlight path changes.

Monitoring
- Track computation latency, queue depth, and explain request rates.
- Alert on repeated timeouts or inconsistent results.

Related references
- architecture/reachability-lattice.md
- architecture/reachability-evidence.md
- operations/score-proofs.md
- docs/operations/reachability-runbook.md
- docs/operations/reachability-drift-guide.md