DX OI On-Premise: Why Alarms aren't propagating to Service

[Alessandro K]

Hi All,

 I configured a Service on DX OI On-Premise but the alarms aren't propagating to Service.

 Here is an example:

Service Overview. We can see the Risk is Severe and we can see "0" alarms on Alarms(+ subs) column :


Topology for Service with components impacting the Service:

The alarms on the component:


In Alarm List there are no alarms:


So, why alarms aren't propagating to Service. Do I need to configure something more?

------------------------------
Regards,
Alessandro
------------------------------

[Nestor Falcon Gonzalez]

Hi Alessandro

I would check if all microservices/pods are running properly on your install. I would also check the logs from the service correlation pods:

analyticsjobs, or soacorrelationengine
Check these steps too:
https://techdocs.broadcom.com/us/en/ca-enterprise-software/it-operations-management/digital-operational-intelligence/20-2/troubleshooting-consolidated/troubleshoot-service-analytics.html#concept.dita_429def2b2674edc32e557a6d7ed46e7fd2b35017_IWantToIncreasetheMemorySizeofServiceAnalyticComponents

I would also recommend to open a Broadcom support ticket to further troubleshoot if above steps are not giving a clear clue of the issue

Cheers
Nestor

[Alessandro K]

Hi Nestor,

Well, I think it will be long thread, I hope that you have time. : )

"I would check if all microservices/pods are running properly on your install."

 When I saw at Cluster Management > Services, there are some services look like it.  I think it is strange because there is no status:

At Kubernetes we have the pods running:

But, the state is Terminated:

------------------------------
Regards,
Alessandro
------------------------------

[Nestor Falcon Gonzalez]

Hi Alessandro

The Terminated status inside the pod is normal. As you can see, it means that one of the Init containers (required only during start-up) terminated with exit code 0 -- all good.
Given all microservices appear as running I would rather check the logs of the pods I mentioned. 
And, in parallel, open a support issue so we can track the investigation.

Thanks
Nestor

[Alessandro K]

Hi Nestor,

 Thank you for reply.

 When you say "I would also check the logs from the service correlation pods: analyticsjobs, or soacorrelationengine"

 I could not see the pods:

[root@master-node ~]# kubectl get pods -ndxi | grep analytics
[root@master-node ~]# kubectl get pods -ndxi | grep soacorrelation

The name of them is analyticsjobs and soacorrelationengine?

In parallel I also opened a case with the same title of this thread.


------------------------------
Regards,
Alessandro
------------------------------

[Nestor Falcon Gonzalez]

Hi Alessandro

if you cannot see those pods it could mean they did not start for some reason. Can you check the logs of the associated deployments to see why they were not started? (e.g. lack of memory, OSE node issues...)
Those pods are responsible for alarm correlation/propagation, so that explains why you are seeing the issue.
In any case, with the ticket open, I am sure you will solve this quickly
Cheers
Nestor