workflows-py/examples/dbos

DBOS Durability Examples

Examples of running durable workflows backed by DBOS. DBOS gives you crash recovery, resumable runs, and multi-replica coordination with either SQLite (zero setup) or Postgres.

See packages/llama-agents-dbos/ARCHITECTURE.md for the underlying model.

Examples

File	What it shows
server_quickstart.py	The simplest durable WorkflowServer setup, using SQLite out of the box. Start here.
durable_workflow.py	A looping counter workflow you can interrupt with Ctrl+C and resume with --resume. Demonstrates checkpointing without a server.
server_replicas.py	Two WorkflowServer replicas sharing a Postgres-backed event store. Start a run on replica A, stream events from replica B, interrupt, and resume. Requires Docker (uses docker-compose.yml).
idle_release_demo.py	Shows how long-idle workflows are released from memory and automatically resumed when a new event arrives.
_replica.py	Single-replica server process used internally by server_replicas.py.

Running

# Quickest path — no database required
uv run examples/dbos/server_quickstart.py

# Multi-replica (needs Docker for Postgres)
docker compose -f examples/dbos/docker-compose.yml up -d
uv run examples/dbos/server_replicas.py