I’m considering using Fly.io’s API to start/create temporary machines for individual tasks and then stop/delete them. In the future, I expect to send a high volume of API requests for machine management actions like updates, creation, and stopping. Are there any rate limits or hard limits on the number of API calls I can make, or on the number of times I can create, update, or stop machines?
Not a fly.io team member, but I believe there’s no hard limits on the API itself. However, there are limits on new organisations, such as a maximum number of total machines you can create (I think it’s roughly 50 or so to begin with). This limit can be raised by contacting fly support.
Machine API rate limits apply per-action, per-machine. Currently, the limit is 1 request, per second, per action—short-term burstable up to 3/r/s/a for everything except the Get Machine call—that one is 5 req/s, burstable up to 10 req/s.
Edit: we now publish Machine API rate limits here.