Timeout after no activity

I’m seeing a timeout after I haven’t interacted with my backend for a certain amount of time. Both in the GUI and also via the GraphQL API. Is there a setting that is autoscaling my backend after no activity? This results in a 30+s timeout on the api. In the GUI I see all sorts of errors such as No Schema Loaded, Wrong API Key, etc. If I just wait and refresh a bunch of times, the backend comes back online

Thanks,
Nick

1 Like

Hey Nick,

What you’re experiencing is a common problem in all cloud services ( Cold start (computing) - Wikipedia ).

A workaround to cold starts is to ensure that you’re receiving atleast one cloud call per 1-2 hours. You can do this in NodeJs by using the setInterval method to make a dummy call after every hour or so

Hi Kaustubh,

I disagree this is a common problem in “all” services. You namely find this with Lambda serverless style functions although you can now configure without cold starts. The problem is how you designed your service. I have cloud services designed on AWS EKS + ALB and it does not encounter any cold starts from our users.

Why does your database have a cold start problem? I’m paying for a backend to stay alive and running and you shut it down to save on your costs?

1 Like