Source record

Published by martinfowler.com (Exploring Gen AI series, incl. Böckeler) (T1 source). The original is at https://martinfowler.com/articles/durable-execution-agents.html.

Pipeline notes

The summary and note below are generated by the signal pipeline — they are Beyond Desk’s reading, not quotations from the source.

The article explains durable execution for long-horizon agents: runtimes persist external side effects, replay programs after crashes or network failures, skip completed steps, and reconstruct state. It separates runtime guarantees from harness design and examines multi-day workflows involving people and external systems.

It gives architects a concrete model for assigning reliability guarantees between the runtime and the agent harness when building multi-day automations.

Cited by

No citations on record.