Filter by State of the app

[rmdes]

When you start hosting 100+ apps it becomes difficult even with the tag system or by filtering by domain to see what apps are running, stopped, crashed etc..

Right now the solution is to scroll down or filter by tag/domain and make sure all apps are running smooth.

but what if we could have another filter on the side of tag/domain for the "state of the app"
Running, Stopped, Error, would be perfect, so that admins could quickly filter apps that need attention, no matter what tag or domain they are under.

[girish]

I like this idea as well Often when logging into customer cloudron's I want to filter by Errored state.

[rmdes]

@girish Exactly and I would even push the idea further : what about having some kind of reporting with this only to admin or super admins ?
right now if one app enters into some error state, there is nothing to get an alert unless one's login and verify the entire dashboard, I think even a simple email would work and if there is a link inside the email pointing directly to the app in question, that would be really neat !

[mehdi]

@rmdes said in Filter by State of the app:

right now if one app enters into some error state, there is nothing to get an alert unless one's login and verify the entire dashboard, I think even a simple email would work

There already are emails sent to admins when an app enters error state / when it is back online

[rmdes]

@mehdi Yup for offline/online/OOM but for not for state of the app with errors ?

[mehdi]

@rmdes I do not understand. When an app is in error state, it is considered offline by cloudron health check, so you get the email. Or am I missing something?

[msbt]

@mehdi there were/are cases in which no email is being sent, when internal healthchecks are positive but the app is broken (either working or non-working, e.g. bitwarden, failed postgres migration in nextcloud/surfer)

[rmdes]

@mehdi I should have kept a copy of a bunch of situations I found like this, no emails received and the app being in a weird state with errors and unable to finish the restart process..

[rmdes]

@mehdi this is my use case : I have 20 minutes for maintenance, I want to login into one cloudron, filter app by states, take care of stalled or error states and voila
get into another cloudron and do the same without having to scroll to check if all apps/domains are running fine..

I'm not very fond of more emails
maybe I should write a simple cloudron cli cron-job that restart apps if stuck in error states
100% of the time those app should be running and 100% of the time restarting the app fix whatever stalled the task in the first place..

I guess that's where having alerts via api/webhooks could be a better way to automate this kind of maintenance tasks when handling hundreds of apps

[robi]

@rmdes
I posted about this in another thread.. I wanted a status at the top of the Admin panel that shows how many are running, how many are not and have those numbers be clickable as a filter to list the ones affected.

Since we're all about automation here, we don't really need a lot of new UI for things that should be automated. Restarts are automatable, it just gives up too soon during an error/timeout/whatever happened.

Adding a backoff timer for App restart will resolve 99% of issues without manual intervention at all.

There's your 20m back.

For those apps that fail to restart after 3 tries, a more reliable notification would be good.

[girish]

@nebulon has added a simple running/not running state filter for the next release. We will iterate on this as we go.

[ruihildt]

@girish It would good to have a separated state for "not responding", which is an error state, and those stopped, which are by choice.

[robi]

@ruihildt said in Filter by State of the app:

@girish It would good to have a separated state for "not responding", which is an error state, and those stopped, which are by choice.

Yes, this was part of my ask and mockup..

[rmdes]

@girish Yay thanks a lot !