Server randomly down. No Errors logs/console. Redeploy 3x in 20 mins. Server started again no clear reason why

Redeploy finished and the server is up now, this is not a solution. On console are no errors so we cant know what the problem is. The endpoints do not fail.

As far as we see qovery is using kubernetes for deployment and running applcations, but we don`t have access there to fix that, also if we had access, redeployment of it is only choise. In this case we need to know what cause issue with backend server down. If its requirement to up resources its not a problem we can do it in moment, but we need to know it. Qovery support should provide more clarification on this.

What info do you need to help us.

Hey @David ,

Can you provide your app console link please?
Why don’t you have access to kubernetes? You can following this guide How to connect to your EKS cluster with kubectl | Qovery

Cheers,

Hello @bchastanier you mean this link? Qovery Web Console | Deploy and Manage On-demand Environments on AWS, Remarkably Fast

About kubernetes, main idea was that we don`t need to access it. Is there any more logs otherwise deployment and app logs, to check whether the problem is

Ok, I understand.
I don’t see any restarts nor health issue from k8s. How did you saw your app was redeploy or your server restarted?
Do you have anything on your side, like graphs or so to help me having a closer look?

Redeploy was done from our side, as app gone into pending mode, and no backend endpoints responded, first hang today was mentioned on [5:27 PM Berlin time], next on [6:04 PM Berlin time]

So the process was next, we use app, in some time its stop responding and all calls was in pending and after that 504 timeout error, we redeploy app, its start working and fails again, next redeploy and working again.

This also happend few times before in past, but we was under impression that heavy scripts of data migration caused this, but todays there was nothing heavy, regular app working in test mode, few peoples using

Ah ! You proceed to data migration before running your app? Usually, you should have little to no down time since we pop the new container and start sending traffic to it as soon as its liveness is ready.

Did you configure your app probes? It can be done via API right now and will be available in our console V3 very soon :slight_smile:

You can configure probes to be OK once migration has been done.

No, it was internal script converting data from one table to others, today there wasn`t any migrations or something like that. Why it goes to hang, could you please provide reason?

Well, that’s a theory since I have nothing ATM to confirm that. But basically, if your app is running but migration is ongoing, it means that app status is failing, and from Kubernetes point of you, if you app is failing for some time, they it kills the pod and deployment fails.

In order to move forward with your issue, you can do the following:

  1. Activate DEBUG / TRACE mode on your app
  2. If not enough, try an APM like NewRelic or Datadog on your app so we can have more information

I don’t see anything weird from Kubernetes, so we need to have more input to move on.

Cheers