Skip to main content
Method Cloud is on only for Method’s own organization for now. For other organizations, method publish --cloud publishes the version, keeps the runs on your workers, and says that Method Cloud is not on.
A production run starts the same way in both places: your app calls method.run(...) or POST /v1/runs. The placement of a Method decides who runs it: Each Method has one placement for each environment (default production).

​
Turn it on

--cloud publishes the version and then sets the placement. --workers sets it back. --env NAME selects another environment. You can also change the placement on the Method’s page in the dashboard.

​
How a cloud run works

  1. Your app starts a run. The server sends it to the cloud runner.
  2. The container gets a key that can use only that run. It runs method worker for that run, with the same runtime as on your computer.
  3. The worker downloads the version by ID, prepares its dependencies, reads the declared secrets from the organization’s secret store for that environment, runs the steps, and sends the run record to the dashboard.
  4. A run can take up to 2 hours. A run whose container stops is queued again and resumes, with the finished steps kept, until its attempts are used.
Your app reads the result with GET /v1/runs/RUN_ID (method.run() waits for it), or gets a signed webhook. Method Cloud is in the Startup plan. See Pricing.