# Reliability

Making agents survive the real world.

## Pages

- [Resilience](resilience.md) — Retry, fallback, health state, circuit breaker.
- [Error Handling](error-handling.md) — Error types, propagation, recovery.
- [Long-Running Runs](long-running-runs.md) — Checkpoint, resume, idempotent retry, runtime cost ceiling for hour-scale executions.
- [Schema Identity](schema-identity.md) — Deterministic schema hashes across LLM providers.