mcpField
M
modal-web-scheduling-knowledge

modal-web-scheduling-knowledge

github.com/josiahsiegelCloud & DevOps5972 installsUpdated 19w ago
Repository

About modal-web-scheduling-knowledge

A reference skill for developers building Modal HTTP endpoints and scheduled jobs. It helps route implementation choices, local testing, deployment, and debugging without covering Modal compute or storage.

Key features

  1. 01HTTP endpoint routing guidance
  2. 02Cron and interval scheduling
  3. 03Local serve and deployment workflow
  4. 04Retries, timeouts, and idempotency guidance
  5. 05Deployment and log debugging checklist

Use cases

  1. 01Choosing an endpoint decorator for an API
  2. 02Setting up recurring ETL jobs
  3. 03Testing and deploying a Modal app
  4. 04Diagnosing schedule or endpoint configuration issues

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/josiahsiegel/claude-plugin-marketplace --skill modal-web-scheduling-knowledge

After installing, type/modal-web-scheduling-knowledgein Claude Code to confirm it loaded.

Details

From the project README

Modal Web and Scheduling Knowledge

Use this skill for Modal web endpoints and scheduled jobs. For GPU compute, autoscaling, or storage primitives, use the focused Modal compute or storage skills.

Endpoint routing

  • Use @modal.fastapi_endpoint() for simple function-backed HTTP APIs.
  • Use @modal.asgi_app() for full FastAPI/Starlette apps, WebSockets, middleware, and routing.
  • Use @modal.wsgi_app() for WSGI frameworks.
  • Add @modal.concurrent for high-throughput endpoints that can safely handle concurrent inputs per container.
  • Test locally with modal serve, then deploy with modal deploy.

Scheduling routing

  • Use modal.Cron(...) for calendar schedules and explicit timezones.
  • Use modal.Period(...) for interval-based execution.
  • Schedules run from deployed apps; modal run is only for manual testing.
  • Add retries, timeouts, and idempotency for scheduled ETL or recurring jobs.

Debug checklist

  1. Confirm the app is deployed when testing schedules.
  2. Check endpoint logs with bounded log streaming.
  3. Validate secrets and environment names before deployment.
  4. Exercise a minimal request or manual schedule trigger before scaling traffic.

Comments

Markdown supported