2 posts tagged with "production"

A working agent on your laptop is not a deployed agent. The gap is mostly boring: a Dockerfile, secrets, a database, autoscaling, observability. Done well, it is one afternoon. Done sloppily, it is a series of 3am pages.

Here is the path from compiled graph to production on AWS. Docker image, ECS Fargate, RDS, secrets, scaling.

The "hello world" agent works on day one. The "we have paying customers and a pager" agent works on day 200. The gap is operational: observability, retries, idempotency, graceful shutdown.

These are the patterns we see in production Python agent codebases, and the failure modes they prevent.