Service Alarm During Maintenance Windows

[kasirgadev]

Hello. I'm just a new dynatrace learner and i've a question.

If i take a host in a maintenance windows, does that host create alarms? i guess not. But i wonder if a service in that host creates an alarm or not? 

[radek_jasinski · Accepted solution]

Hi @kasirgadev 

It all depends on how you configure Maintenace Window.
You have 3 modes:

• Detect problems and alert: Dynatrace automatically detects and reports all problems as usual and displays a maintenance window icon on each problem that is detected during the maintenance window.
• Detect problems but don't alert: Dynatrace detects problems but does not send out alerts for them. Each problem is listed on the Problems page with a maintenance window icon.
• Disable problem detection: Dynatrace does not detect problems or send out alerts for them. Problems that occur during scheduled maintenance windows are not included on the Problems page and no alerts are sent out.

Radek

[kadirhan]

Hello ,

Actually, we are wondering about this here. We selected the alarm generation option and added a new rule. However, we only added host to the definition in the rule.

Rule Create an alarm for service alarms on this host? Or does the rule affect alarms in all entities to which the host is connected?

We only added the host, do we also need to add service entities to the rule? Or does it not create those alarms by saying it is on this host in the service?

Best Regards

[jtclark]

I'm facing the same issue - and no, scheduling for the HOST entity does not include the child entities dependent on the HOST. Each entity to be suppressed/actions need to be in the MW. There are a couple or RFE's out there wanting a toggle on the MW page to include child/dependent entities.

[AntonPineiro]

Hi,

I would suggest you using auto tagging rules for propagate tag to child entities. And create maintenace window base on tag.

Best regards

[jtclark]

Yes, that's the recommendation I keep seeing, but it is just another layer of administrative and configuration overhead that can be avoided with a simple change to the MW interface to simply include dependent service and process instances. The are many posts and product suggestions for this in this desire, and the logic is something easy to configure as Tag rules already has this ability.

[kadirhan]

Hello,

You say to do it with tag, but we want to bring the services into the maintenance window for specific hosts.

For example, there are 6 servers under one service. We only want to bring 3 of those 6 servers into the maintenance window. So we need to do it instance based. However, if we do this using tags, the whole service will be taken into the maintenance window and this will result in us not receiving critical service alerts from the hosts that need to be working.

How do we solve this problem? How can we say Get Service Data (Service Instance) to Maintenance Window for this list of hosts?

Best Regards,

Kadirhan Cekmez.

[jtclark]

@kadirhan Exactly why I'm not a fan of the Tags idea. For now, the only solution I can think of is to replace the MW UI with a custom app that I can enter the host and then call the Monitored Entities API to get the "toRelationships" object, and parse each array under it for disks, services, process group instances, etc., but leave the Many to Many entities like Process Groups out since they can exist on multiple hosts.