Managed VMs: Are connections drained before the weekly restart?

[Jon Keys]

I was reading about the weekly restarts in the Managed VMs service and I'm just curious how the restarts are handled.

For example, if the managed vm is handling an active request and the machine gets restarted will the request get lost in the ether? I would expect there to be some sort of guard to protect against this scenario (e.g. connection draining or maybe a hook to notify the app of a pending shutdown) but I can't seem to find anything in the docs.

Of course my app should be resilient to restarts but it would be better to avoid problems if possible.

[Josh Moyer]

Hi John,

Similar to regular backend shutdowns, a Managed VM application is sent an /_ah/stop request to give it a chance to do clean up.  Existing requests will have some time to complete, but they may get interrupted if they are long running.  I have issued a request to have the documentation that you cited updated with this information.

Kind Regards,

Josh Moyer

Google Cloud Platform Support

--
You received this message because you are subscribed to the Google Groups "Google App Engine" group.
To unsubscribe from this group and stop receiving emails from it, send an email to google-appengi...@googlegroups.com.
To post to this group, send email to google-a...@googlegroups.com.
Visit this group at http://groups.google.com/group/google-appengine.
To view this discussion on the web visit https://groups.google.com/d/msgid/google-appengine/f3d17923-5ff3-4041-819c-606bfabf84c2%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.