Hi Fellow users,

Do you encounter slow down in Cost transparency?

Recently we encountered a lot of slow down (10 minutes to open the dev instance and a show a repport)

it seems that the apptio team is not doing any change at the moment and most of the time we have to request a reboot to go back to our usual pace.

thanks for your feedback

Hi wilfrid Adrien, 
I think there must be a specific reason for the . Please contact your . He or she can provide details on the live performance and also concerning the calculations running in the background.

Best regards, Fabian

Hi wilfrid Adrien,

We used to have the same problem. We are on R12 platform and I have a number of pointers for you with regards to R12 some of which can be actioned immediately, others will require configuration work.:

1) If you have multiple branches, try and remove them, each branch takes resource, slowing down your instance.

2) If you have multiple users checking stuff in and out, then try and get your users to only check in once at the end of the day. This will then mean calculations will run out of normal working hours so slow down won't be as obvious

3) Get Apptio support to give you a full rundown report (they do this for every customer once a quarter as far as I understand) this will indicate any poorly performing tables and reports.

4) Look at your modeled tables - do they have a specified unique row identifier column (or columns) if not, try and set it, do not rely on Apptio's internal identifier.

5) If your R12 instance was created from and r11 upgrade, check all your allocations and make sure they are not using the old r11 inference joins. There are documents in TBM connect on how to correct this.

6) Finally review your model and look at the master table granularity - can it be reduced? see my blog - The Pain Of The Grain - An ETBMA Case Study

Regards,

Chris

Hi @Christopher Wilson, hi wilfrid Adrien,

our slowness was mainly caused by the trickle process at the end of a

To trickle means "finishing" the calculaiton, when the build is trickling to the workspace nodes. The calculation nodes were not affected.

There were a number of long running threads in the workspace nodes that slowed down the trickling process. The reason for that was the missing object identifier in some of our model objects within the model step.

We solved this by checking each model object and set an object identifier. Why some of the identifiers were not set before, we don't know. This helped to reduce the pure tickle time from over 4 hours to less than 30 minutes.

Furthermore we used to pause the STG calculations during the day, because STG calcs sometimes slowed down the system. Apptio re-enabled calcs at the evening. And cancelling the STG calcs was not an option, because this can lead to a corrupted environment, which can only be resolved by a environent full reset & cache clear. 

BR, Fabian

HI @Fabian Woldsen,

You have made some good points. My item 4 should read Object Identifier rather than row identifier - I'm an ex database developer and still get tied up in that 'language'.

I must admit pausing STG calculations is a new one on me. Is there a way to configure this in TBM Studio or is it an Apptio Support request?

Thanks,

Chris

Hi @Christopher Wilson,
currently STG calc can be paused by Apptio support only.

But this is a planned feature for an upcoming release. I think your CSM should know about the Apptio internal feature request ticket.

Best regards, Fabian

Hi Wilfried

We have currently also slowness on our DEV environmment (R12), it seems on our side to be link to long DEV Calculation "Finishing" step (where a copy of the Calculation result is performed).

If you see also a step finishing longer than 10 minutes for each DEV calc then you may have the same issue.

APPTIO will provide us a specific patch , but till now not implemented.

Regards

Rachid

I think this is heavily dependent on your own config and how you're using branches, checked out docunments etc. I would request a model review with the support guys. Might help you diagnose the issues.

Yes, I have as well.

We do experience the occasional slowdown in our R12 Dev environment every now and then - even after being migrated servers. We usually experience this through what we call the "Gray Screen of Thinking" because of how R12 just goes gray for several minutes without us being able to do anything but think about what caused it.

Moving forward, I'm really hoping that this becomes less frequent...