Highlighted
Super Contributor.. twinge Super Contributor..
Super Contributor..
826 views

SLA - Cluster Availability Rule

Jump to solution

Hi all!

 

We have a transaction that runs on 2 PCs and I have created an SLA that calculates the availability with the "cluster availabilty rule".  

The problem is that the cluster availabilty calculated from the SLA is even worse than the worst child and I cannot understand why. My worst case should happen when both PCs fail at the same time. But how can be the cluster availability even lower than this?

 

I attach 2 screenshots, maybe someone can help. 

Please note that only the availability is calculated with the cluster rule (the other KPI are "best child") and that the calculation method of the children is "time based". 

 

Thanks and kind regards!

0 Likes
1 Solution

Accepted Solutions
Micro Focus Expert
Micro Focus Expert

Re: SLA - Cluster Availability Rule

Jump to solution

There are few things you need to understand about Cluster availability rule:

1) it is calculated based on the children KPIs, but within 5 minutes buckets

2) according to your settings, the parent will be considered as available if there is at least 1 child that its KPI's result is at least 100%

3) this means that if the children results are not 100% then that 5 minutes bucket will be considered as not available.

 

in order to better understand the results, you need to look at the children raw data and see where each child failed and how many overlappings you had. This can explain how the parent will have less availability than its children.

0 Likes
3 Replies
Micro Focus Expert
Micro Focus Expert

Re: SLA - Cluster Availability Rule

Jump to solution

There are few things you need to understand about Cluster availability rule:

1) it is calculated based on the children KPIs, but within 5 minutes buckets

2) according to your settings, the parent will be considered as available if there is at least 1 child that its KPI's result is at least 100%

3) this means that if the children results are not 100% then that 5 minutes bucket will be considered as not available.

 

in order to better understand the results, you need to look at the children raw data and see where each child failed and how many overlappings you had. This can explain how the parent will have less availability than its children.

0 Likes
Super Contributor.. twinge Super Contributor..
Super Contributor..

Re: SLA - Cluster Availability Rule

Jump to solution

Hi Asaf, thanks for your explanation!

 

I attach 2 more screenshots, just to completely understand. 

 

From screenshot03 you see the only fails of the day of the choosen transaction. The second PC has 1 fail more, so it is completely clear why the CI Summary shows a lower availability of this PC for the choosen day (screenshot04). 

 

You can also see the cluster availability of the day is lower then the worst child (PC2), should I assume that there could be an overlapping of the 1st PC (at time 8:25) that matchs the second PC and/or for example an overlapping of both PCs at time 9:00, and this harms the cluster day availability, bringing it under the worst child?

 

Do you suggest to remove any offset to avoid this problem (there is a 120 offset now)?

 

Thanks again for your precious support!

0 Likes
Micro Focus Expert
Micro Focus Expert

Re: SLA - Cluster Availability Rule

Jump to solution
there could be overlapping. you need to look at the raw data report and see the exact time stamps of the failures.
0 Likes
The opinions expressed above are the personal opinions of the authors, not of Micro Focus. By using this site, you accept the Terms of Use and Rules of Participation. Certain versions of content ("Material") accessible here may contain branding from Hewlett-Packard Company (now HP Inc.) and Hewlett Packard Enterprise Company. As of September 1, 2017, the Material is now offered by Micro Focus, a separately owned and operated company. Any reference to the HP and Hewlett Packard Enterprise/HPE marks is historical in nature, and the HP and Hewlett Packard Enterprise/HPE marks are the property of their respective owners.