Prometheus - monitoring & time series database

[eriktad]

@girish Yes but I only find basic http auth on Prometheus docs, so what can you suggest?

[eriktad]

@girish is it normal practice for us?

[girish]

@eriktad Yes. You can do something like https://git.cloudron.io/cloudron/simple-torrent-app/-/blob/master/apache/cloud-torrent.conf . That apache config sets up the app for authentication via Cloudron LDAP .

[eriktad]

@girish cool thx

[eriktad]

@girish I tried this example, but somehow apache does not start for me, if you can help, I will give the more information about this

[girish]

@eriktad Happy to help, let me know what error you are facing.

[eriktad]

@girish if I try to run start.sh manually it gives me

[eriktad]

@girish So idk what happens I try to do it like another example I found in GitLab

[eriktad]

@girish Ye problem founded. #!/bin/sh -> #!/bin/bash

[eriktad]

@girish Prometheus app is done. Can you review it and tell me what to need to do next?

[eriktad]

@girish https://git.cloudron.io/erik/alertmanager-app and Alertmanager app

[girish]

@eriktad That's great progress! Yes, I can take a look next week. I have marked this app as WIP.

[girish]

@eriktad said in Prometheus - monitoring & time series database:

@girish https://git.cloudron.io/erik/alertmanager-app and Alertmanager app

I don't have much experience with Promotheus. Does it make sense to merge alertmanager into the main app itself or do you think it should be separate apps?

If you need help deciding, we can take a guess from how they are usually deployed? Are they in separate domains or as subpaths etc.

[eriktad]

@girish Yeh, I thought about that but idk, they run on other ports and have other dashboards, but altermanager only alerting Prometheus massages, I can try to merge, think it is good idea, but we need to understand if user no need alerts why we setup it with Prometheus.

[marcusquinn]

@eriktad They make sense as separate Apps.

I'd suggest the App naming would make it clearer, so:

• Prometheus Server
• Prometheus Alertmanager
• Thanos (Single-Node)

For interest, like Statping, in mission-critical setups it makes sense to have Prometheus running on a separate Cloudron instance for continuity of data-gathering throughout the ups & downs of the services being monitored.

[girish]

@eriktad Thanks for packaging the app. I have published the app as unstable currently. https://git.cloudron.io/cloudron/prometheus-app/ is the repo. I removed the apache authentication bits because I have now merged the "auth wall" feature in Cloudron 6. With that feature, we can just add proxyAuth to the manifest and we will immediately get a login screen (so we don't need to put these things as part of the app).

[girish]

@eriktad Can you do similar changes for the alertmanager repo as well? I will publish that as well.

BTW, I also put you in the app dev group. You might have to go to your profile so that the badge displays.

[girish]

I moved the repo to https://git.cloudron.io/cloudron/prometheus-server-app per @marcusquinn 's recommendation

[eriktad]

@girish yes thx.

[eriktad]

@girish alertmanager new version pushed to my repo.