Feedback Thread: Observability experimental feature

[jamescarppe]

In Portainer Business Edition 2.32.0 STS we introduced a new feature - Observability. This feature is experimental and in the first iteration, and we are keen to gather feedback on it via this thread.

• What do you like about the feature?

• What do you not like about the feature?

• What improvements / changes would you like to see for the feature in the future?

• Any other feedback around the feature specifically?

The more feedback we can get the better we can make the feature in the future!

[alexislefebvre]

Noob question: how do we enable experimental features or this feature?

[jamescarppe]

In 2.33 LTS, under Settings, you'll find an Experimental features section. Here you can toggle Observability on. You can then access it in the left hand menu.

In 2.34 STS and above, this has moved out of experimental and can be toggled in Settings, Additional functionality.

[alexislefebvre]

In 2.33 LTS, under Settings, you'll find an Experimental features section. Here you can toggle Observability on. You can then access it in the left hand menu.

In 2.34 STS and above, this has moved out of experimental and can be toggled in Settings, Additional functionality.

Thanks for your answer. I use “Community Edition 2.33.1 LTS” and there is no “Experimental features” section.

---

Update: my bad, I missed something from the announcement: “In Portainer Business Edition 2.32.0 STS we introduced a new feature”. So it won't appear on my Community Edition.

[webysther]

I think the better observability is a better notification system, the current one don't show if the stack stuck or if the image is being downloaded.

The mostly rules are too specific and observability need to give a better overwall view not about small things that happens sometimes.

For instance:

• Show the detail about stack deploy like time to download image
• Link in notification for Observability for more details or the full stack trace
• Allow in Observability the full stack trace all around
• Decent overview dashboard for the rules

I have to build myself the real observability for the stacks years ago, which works because I only use naming convention and labels:

Why not starting using labels and other thinhs available to do this?

Group by stack (or project):

Group by type (using just a stupid name pattern):

[webysther]

And before starting talk about this, just finish the basic stuff:

https://github.com/orgs/portainer/discussions/9344

How I will oberve something which is incomplete? Make no sense.

[Stealth-ip]

Please add a Telegram option. Since Telegram requires both chat_id and text within the payload, a regular webhook (without a custom JSON body) won’t work.

[Finkregh]

Alerting via webhooks

The documentation shows zero details, is it POST? GET? How would I pass additional parameters? E.g. for https://pushover.net/api

Alertmanager?

Reading from the logs, you seem to be using alertmanager:

2025/10/14 04:34PM ERR github.com/portainer/portainer-ee/api/internal/metrics/alerts.go:361 > Failed to send alert to Alertmanager | error="Get \"http://localhost:9093/api/v2/status\": dial tcp [::1]:9093: connect: connection refused"

Why not just expose the alertmanager/prometheus/metrics and/or

• alerts config in a file, which could be consumed by an existing alertmanager/victoriametrics/...
• expose /metrics directly

When letting people change the configs directly, you could have a big banner in the UI telling them, that they are now on their own for that functionality. But powerusers would very much enjoy that :)