Alerting for Kubernetes-oriented metrics

[davide_diana2]

Are we able to gather Kubernetes-oriented metrics in order to rise alarms based on them? Customer is asking to get a notification when Kubernetes says the container / service is not correctly working (see screenshot - pod is running but not ready)

Are we able to provide these kind of KPI?

[daniela_rabiser]

Hi @Davide D.,

we plan to add Kubernetes-specific alerting, e.g., to notify about stuck and pending pods. In general, we plan to add Kubernetes related anomaly detection capabilities for Dynatrace. This is on our roadmap, however, I'm not yet able to provide an ETA.

Hope this helps for now.

Best,
Daniela

[davide_diana2]

Hi @Daniela R.,

thanks for the insights! It would be great to get these kind of kuberntes-specific metric soon.

In order to give you other use-cases, customer would like to get notifications related to resource consumption of each pods compared to the limit they set on them:

[...]

resources:

limits:

cpu: 1

memory: 1.5Gi

[...]

Thank you,

[kamal_kyrala]

I would also love to see this in AppMon

[baskaran_selva2]

Guys, Any updates for Kubernetes specific metrics, it would be very helpful if we get sooner

[baskaran_selva2]

@Daniela R - Do you have Any update on Kubernetes specific metrics, if we get sooner it would be very helpful. I remember we discussed at perform conference, and you said it will be released sooner, the reason we are pushing that currently, we are using two monitoring agents kubernetes clusters and another one we specifically using for kubenetes pod availability and related those metrics

[baskaran_selva2]

@Daniela R - Do you have Any update on Kubernetes specific metrics, if we get sooner it would be very helpful. I remember we discussed at perform conference, and you said it will be released sooner, the reason we are pushing that currently, we are using two monitoring agents kubernetes clusters and another one we specifically using for kubenetes pod availability and related those metrics

[alois_mayr]

Which metrics are you interested in? We are about release a first version of an plugin that captures metrics from Kubernetes. It's going to be published on GitHub (https://github.com/dynatrace-innovationlab/k8s_activegateplugin), however, the repo is private at the moment and feel free to use, contribute and provide feedback once we make the repo public (next week most likely).

[jan_palic]

Hi,

is a ActiveGate on windows needed or do we get a positive test with a linux based ActiveGate?

Best,
Jan

[mallempati_sree]

hi Alois,

Am getting 404 for the Git link, has it been removed ?

[drors]

Joining former members on this issue. Creating alerts from k8s metrics is highly needed. We can track events when a pod is restarting but cannot create an alert from this data. what is the current ETA fro that? Thanks, Dror

[STiAT]

We'd require information on stuck and pending pods too to create alerts from it. I'd be interested in an ETA or if it actually is still on the roadmap (it's been quite some time).

[enricostahn]

Any update on this topic?

[Stephen1_Lee]

This is urgently needed as well as the ability to get metrics based on each namespace.

[rodolfoalbuquer]

We'd also like the ability to get metrics on node events like, NodeNotReady or NodeHasSufficientMemory transition and so on. We can already see the events on the Kubernetes integration section but can't alert on them.

Any updates on the topic?

[tracyfields198]

Hello everyone,

Are there any updates for kubernetes specific metrics.

We are waiting for the updates so eagerly.

Please keep us posted on the updates.

[LNarayanan_]

Hello Community,

Do we have any update on adding Kubernetes-specific metrics? It is much required for setting up alerts.

I would be needing metrics such as below:

1. Metric to alert when a Pod or Node Autoscales
2. Metric based on Node or Pod memory and CPU in Percentage(%) (Currently we have for millicores(unit))

As shown below we don't have these metrics for Pod and Node memory and CPU in % but we have metrics in % for Cluster.

Metrics for Cluster CPU in %:

Ref:

https://www.dynatrace.com/support/help/how-to-use-dynatrace/metrics/built-in-metrics

Thanks in Advance

[MaciejNeumann · Accepted solution]

Hello everyone,

Here is the link to the documentation article on all the built-in metrics for Kubernetes:
Built-in metrics - Kubernetes 
And the new list, with all the built-in metrics on Grail for Kubernetes:
Built-in metrics on Grail - Kubernetes 

Here you can also find a Feedback channel from one of our Product Managers responsible for the Containers topic:
Introducing easy alerting on common Kubernetes platform issues 
If you lack anything that you need for the Kubernetes-focused alerting, just comment in there to let us know 🙂