Re: SLA Time Based Calculation - Month vs. Quarter (162 Views)
Reply
Frequent Advisor
twinge
Posts: 53
Registered: ‎06-25-2012
Message 1 of 4 (206 Views)
Accepted Solution

SLA Time Based Calculation - Month vs. Quarter

[ Edited ]

Hi All,

 

Maybe I am missing the point but is it better to ask. 

One of our time based SLA shows following situation (same for all SLAs, this is only one of them):

Availability Month October:
1) 99,967%
2) 99,944%
Group Average: 99,956% - correct

 

Availability Month November:
1) 100%
2) 99,519%
Group Average: 99,760% - correct

 

Availability Month December:
1) 100
2) 100
Group Average: 100 - correct

 

Average Availability Quarter Oct-Dec (CI Summary/CI Status):
1) 100%
2) 99,998%

Group Average: 99,999% - seems to be correct but...

 

...how can we have an average availability for 1) of 100% if the 3 Months were. 99,967%, 100% and 100%?

Same for 2). How can be the average availability 99,998% if we have for each month 99,944%, 99,519% and 100% ?

 

Changing the calculation method to sample based, we have a value for the average quarter availability that is near to the truth:
1) 99,989%
2) 99,827%

Group Average: 99,908%

 

Here a screenshot  (SLA V2 is Time Based - SLA-V3 is exactly the same but Sample Based).

Is there an explanation?

 

Thanks for the help! 

Please use plain text.
Frequent Advisor
twinge
Posts: 53
Registered: ‎06-25-2012
Message 2 of 4 (204 Views)

Re: SLA Time Based Calculation - Month vs. Quarter

[ Edited ]

Here the screenshot again, in the post is too small. Thanks.

Please use plain text.
HP Expert
Asaf Shechter
Posts: 1,210
Registered: ‎12-09-2008
Message 3 of 4 (171 Views)

Re: SLA Time Based Calculation - Month vs. Quarter

it is not easy to answer w/o looking at the system.
I also see that you use custom Calendar, which can have impact on the calculation.
The 1st point to check is why 1 location gives 100% availability in time-based and 99.x% in the sample based. make sure both of your SLAs start at the same time and you recalculate them from the same time.
for more info, I suggest to open support case.
Please use plain text.
Frequent Advisor
twinge
Posts: 53
Registered: ‎06-25-2012
Message 4 of 4 (162 Views)

Re: SLA Time Based Calculation - Month vs. Quarter

Thanks Asaf for your answer. 

To check the start point and the recalculation point of the SLA was a good idea. 

The V3 SLA (Sample Based) was a Clone but started not at the same time as the V2 SLA (Time Based). 

 

What I did:

Cloned the V3 SLA (Sample Based) into V4 SLA (Time Based) with the same starting point and recalculate them.

 

So I had:

V2 SLA - Time based - old SLA

V3 SLA - Sample based - new sla starting from a later date

V4 SLA - Time based - new sla starting from the same date of V3

 

Results:

The monthly CI reports of V4 SLA displayed other values for the monthly availability then the V2 SLA. The quarter report of the time based V4 SLA was also correct.

 

Please note:

Some hours before the clone and recalc., I installed the patch for BSM 9.23 and I do not know is this solved some SLA issues of BSM 9.22 

 

All SLAs (V2, V3 and V4) had the same amount of samples (before and after the installation of the patch)

 

The final check was to recalculate the V2 SLA after the patch installation and it suddenly showed the same values of V4 SLA. So maybe the patch solved really some SLA issues of BSM 9.22

 

Thanks again and kind regards!

Please use plain text.
The opinions expressed above are the personal opinions of the authors, not of HP. By using this site, you accept the Terms of Use and Rules of Participation