The Render documentation says that there is a timeout for the healthcheck to respond with a code between 200 and 399 but doesn’t mention what that timeout value is. One of my services fails periodically, and I think it’s because database migrations are taking too long to complete.
Is the timeout duration documented anywhere and is it possible to configure a longer timeout for services that have significant work to perform during initialization even once the build is successful?
Render will wait 15 minutes before marking a deploy as timed out. We’ll push an update to the docs to make that clear.
We don’t currently allow that duration to be configured. If you need to run something for a longer period of time, I’d suggest using Render jobs instead. It allows you run a task to completion (e.g.
migrate) using a copy of your existing service.
Our deploys are failing after 3-5 minutes so given what you’ve said, the timeout probably isn’t the issue.
It’s an instance of Strapi and we’ve had a really difficult time getting the application to deploy consistently on Render. Sometimes Render shows the deploys has failed but logs keep coming through for a few more minutes, leading my to believe that Render timed out. This could be a red herring if logs are delayed but the dashboard is already responding to some other signal that the deploy has failed.
Thank you for the quick feedback–I’ll keep looking into the Strapi issues.