Build durable agents with Temporal and Lakebase

Summary

Long-running personal-loan underwriting agents must preserve completed work, resume after worker failures, tolerate repeated tool calls, wait for reviewers, expose operational state, and retain an audit trail. The reference implementation combines Temporal durable Workflows and Activities with Lakebase Postgres, using Temporal Event History for replay and Lakebase schemas for run status, evidence, review state, messages, and metrics. Deterministic identifiers, constraints, guarded updates, and Postgres upserts make repeated Activity attempts target the same logical records, while Signals carry human decisions and long waits do not occupy workers. A Unity Catalog synced table supplies governed underwriting policy, and optional Change Data Feed publishes operational changes to Delta history tables. The implementation includes a React/FastAPI interface, mocked deterministic providers, OAuth credential refresh, and 21 passing tests, but does not validate production lending, compliance, security, scale, or an observed end-to-end feed.

Takeaways

1. 1

   Temporal Event History lets a replacement worker replay recorded Activity results, timers, and Signals; an Activity whose completion was not recorded can be attempted again.

2. 2

   Lakebase stores application-facing run status, evidence, messages, review state, tool calls, events, and metrics separately from Temporal’s replay state.

3. 3

   The implementation does not include an observed end-to-end Change Data Feed run, and the feed requires enablement and destination verification in the target Databricks environment.