a year ago
The "serverless" (formerly "sleep" right?) is active in my fastAPI application in Railway. There have been a few occasions this week when a request was not received (or least I didn't see a trace of it in the logs). The container started and then closed, but the request was not queued as far as I can see.
I would prefer not to disable the sleep/serverless feature, so I am considering a recursive call to a dummy /awake endpoint that would take place prior to the actual request to "awaken" the app, and would iterate a predetermined numbered of times (say, 5) while it gets status code 502. Would you say this is a good approach to avoid the problem?
Or are you aware of any issues on your end that might be related?
Thanks, Manuel
<#1006629907067064482>