Host AI Workflow Runner with Flask: Worker Execution and Job State

Summary

Deploy an AI workflow runner built with Flask as a public web service plus a dedicated worker process on Ample. ample plan --write discovers both services and their shared database and writes ample.toml; ample up reconciles the topology: a managed PostgreSQL 16 database, a public web microVM and a private worker microVM (kind worker, no public URL), both reading DATABASE_URL. Verified on Flask: a separately deployed worker service (kind worker, no public URL) sharing a declared managed PostgreSQL database with the web service, consuming a jobs table with row locks and marking jobs done once (worker=done attempts=1). Not separately tested: your job payloads, retry policy and scheduling; treat the AI workflow runner-specific behavior as your application code.

Resource Requirements

• Compute: Apps run in isolated x86_64 Firecracker microVMs that auto-pause when idle and wake on request; sizes are the priced VM sizes.
• Postgres: Managed PostgreSQL 16 runs in its own microVM and is auto-provisioned when an app needs a database and no DATABASE_URL is supplied.

Framework

Flask

Workload

AI Workflow Runner

Execution Status

Ready

Prerequisites

1. A repository with the web app and the worker as separate service directories (the fixture uses apps/web and apps/worker), each building and starting with the documented commands.
2. Both services read DATABASE_URL at runtime and share one jobs table; the worker claims rows using SELECT ... FOR UPDATE SKIP LOCKED and marks them done once.
3. An Ample account token with servers:write and databases:write.

Workflow Steps

1. Plan the topology: Run the planner once. It discovers the web and worker services and writes ample.toml. Exit 2 means it left questions in the manifest; answer them with --answer or by editing the file. (Command: ample plan --write .)
2. Share one database: Keep a single [databases.main] with engine = "postgres" under both services. Python services need an explicit start command in the manifest. (Command: ample plan --offline .)
3. Apply: Reconcile the whole manifest in dependency order: the database first, then both services. It is idempotent and never destructive. (Command: ample up .)
4. Verify: Confirm the worker processed the enqueued job. (Command: ample logs --kind runtime)

Examples

• Flask web + worker fixture: A Flask web app and a worker process sharing one managed PostgreSQL database.

Success Checks

1. Web service responds on its public URL.
2. Worker consumed the enqueued job.

Limitations

1. Verified on the python-3.12 template at s-1vcpu-1gb for both services.
2. The AI workflow runner itself (job payloads, retry policy) was not separately tested; only the worker-queue check is verified.
3. The worker service is private and has no public URL.