Application passing health checks, but not available, and nothing in logs

Hi,

I redeployed my application (name “colours”) yesterday, and have an odd problem. The health checks are passing (the deploy succeeded), but the application is not accessible. I can SSH into it, and I can see it’s running fine (It’s Elixir so I can connect to the running instance after SSHing in). However I cannot access it at the public address. I can ping the address, but curl hangs.

Re-deploying has no effect (it successfully redeploys, but no change)
Scaling to 0 then back to 1 has no effect (same as re-deploying)

The logs are basically empty - they show the application starting up after deploys, but then show nothing (as requests are not reaching the pod).

It is deployed in LHR, and I know there was some maintenance going on yesterday so I am wondering if it is related?

Thanks
Zander

This application is back up now. If anyone sees this, would be interested to know what the cause was.

I think this was a service propagation issue. Your new VM didn’t get registered properly when it booted. I cleared that up and was able to connect.

This is a bug on our end, but it’s worth noting that apps that run 2x instances are less susceptible to these kinds of problems. Single VM apps are more prone to failures like this. We’re improving the reliability of our service discovery so you may not ever notice, but if an app is important you should run 2+ instances.

Thanks for the quick reply. Very good point.