SLA Report for Day & Night

Meta Icinga Events and Community
1

Hello,

We are monitoring branches & DC devices. We are having different SLAs for branches & DC devices. For branches the SLA will be paused during the window 8PM to 8AM but we need to monitor the availability for 24 hrs. For DC devices the SLA clock will be 24*7 running. Is it possible to take the SLA reports of branches only for the time frame 8PM to 8AM. How will we configure the reporting module for SLA.

2

The is a pull request addressing this:

Here is a tutorial explaining how to apply it as a patch: github - How to apply a git patch when given a pull number - Stack Overflow

1 Like
3

Im sorry I don’t see it. why does a Fake downtime in another table change the result of an SLA calculated in the Mysql function idoreports_get_sla_ok_percent? Best regards Nicolas

4

The reporting module depends on idoreports

5

The merge request was closed with a tag won’t fix so this won’t help ): You can see this on Fake downtimes by nmartinii · Pull Request #54 · Icinga/icingaweb2-module-reporting · GitHub
So does anybody have any idea how to process further?
With the current implementation I would expect a lot of false positives with the SLA reports due to maintenance windows

6

I’m also in the situation like above, we have different SLA levels for different objects in Icinga and we would like to reflect that in the SLA reports.
Let’s say a service on a host has SLA level reaching from Mon-Fri and it goes down during the weekend, that would now mean that the downtime during the weekend is included in the calculated SLA which is not correct. Or am I missing something?

7

We had kinda the same issue, but we solved the “false positive not reachable hosts” with scheduled downtimes that we created for all objects that are not in the 24x7 SLA.
Example: For all officeday-hosts we scheduled a downtime for the complete weekend.

8

OK, did you group all “SLA-Level1” hosts/services into a group and then run a downtime schedule towards that group? Sounds like a way forward in my case also.

9

Nope,
we have defined time_periods (in Director it’s configured here: /icingaweb2/director/timeperiods) and they are added to the host as custom vars.
Each host has a custom var called notification_sla, so you can define the “Assign where” of the scheduled downtime with a host.vars.notification_sla = "officedays" or host.vars.notification_sla != 24x7

1 Like
10

That sound much better… :slight_smile: Thanks!

1 Like