Process group downtime per hostname in dashboard

[ALF]

Hello 

I'm looking for a way to visualize in a dashboard the moments when a process group instance is down on a host.

At this time I'm using the following metric query

builtin:tech.generic.processCount:filter(in("dt.entity.process_group_instance",entitySelector("type(process_group_instance),entityName.equals(~"My Process Group Name~")"))):parents:splitBy("dt.entity.host"):min:names:sort(dimension("dt.entity.host.name",ascending))

On a dashboard the hostname disappears when the service is down on the host and shows if it's up.

On the contrary, I would like the hostname to appear on the graphic, only if the process group is KO on the hostname.
Is there a metric query to achieve this ?

I attach a screenshot of my actual dashboard.

Regards

[radek_jasinski · Accepted solution]

Hi,

As far as I can recall, such a mechanism that something appears and disappears (as in the case of the hosts you write about) is not present in Dynatrace dashboards.

You might consider to verify the number of active process instances within a process group on a given host:
builtin:tech.generic.processCount

[MarwanC]

Hi Radek,

I appreciate your reply

"You might consider to verify the number of active process instances within a process group on a given host:
builtin:tech.generic.processCount"

I would like to do that, how does the query look like in a graph or in a single value dashboard,

I used this successfully to look for the CPU of that particular process

builtin:tech.generic.cpu.usage:filter(eq("dt.entity.process_group_instance",PROCESS_GROUP_INSTANCE-98CB935943677AB7)):splitBy("dt.entity.process_group_instance"):max

I want to have that particular PROCESS_GROUP_INSTANCE

I appreciate you support

Marwan...

[MarwanC]

I found the solution by grouping by technology and it works like a charm, I share the code below for anyone who wants to do that.

builtin:tech.generic.processCount:filter(and(or(in("dt.entity.process_group_instance",entitySelector("type(process_group_instance),softwareTechnologies(~"WEBSPHERE~")"))))):splitBy("dt.entity.process_group_instance"):sort(value(auto,descending)):limit(100):lastReal

The only problem still it only shows that processes which are up but not those that are down. When the process is down Dynatrace does not collect and therefore it only shows what is up. This is not very useful.

Any idea?

[MarwanC]

I think I was wrong, it should show all up or down state. I just needed to play around with the time frame. So the above solution is perfect and can be used for any process name or technology. This question can now be closed, I hope is useful for eveybody.