Any changes or updates made on a production system should always be tested, signed off and approved for deployment. With larger TM1 implementations and multiple internal and external TM1 developers working in parallel silos and with different release cycles, it can be hard for the system owners to track the changes status on a given TM1 production environment. It can be even harder when under stress as production issues are raised by users shortly after a deployment did occur.

With the change tracking feature in Pulse, every time an object is changed by either an administrator, a user or a system process, it is tracked and logged in Pulse. With Pulse, Stockland administrators are able to track all changes in the various environments and ensure they are aligned to the planned deployments.

The admin team is able to trace when a change is made, by who and on what component. With Pulse you can even see which lines of the code (Rules/Tis) have been updated. Combined with the Pulse rollback features you can also restore the previous version of a TM1 TI or rule when invalid changes are identified.

“Pulse helps us to track and verify the correct changes have been made in production”

We apply the classic development, test, sign off, deployment delivery mechanism for our TM1 environments. We have a large number of production instances and on occasion have multiple developers working concurrently to deliver updates to various models which can be in the same of different instances. Following a series of releases, we started to receive alerts from Pulse indicating users were not able to log in one of our TM1 instances. Users were locked because of an updated TM1 process which was previously tested and signed off in DEV and UAT but seemed to cause the issue in PROD. Using Pulse’s Change Tracking feature, we were quickly able to determine that a process was incorrectly updated in production by an external contractor as the changes highlighted did not correspond to the defined deployment plan. In effect, the wrong version of the TI had been deployed (Human error) from Test to Production. We could see exactly which lines of the TM1 process were updated , roll back the changes and redeploy the correct TI version to resolve the issue quickly limiting user impacts.

“TM1 is slow !” When TM1 users complain that TM1 is occasionally slow, it’s fair to say that with the native TM1 administration toolkit the analysis required to identify the performance issue root cause is time consuming and cumbersome. How to quickly determine if the performance issue is model design related, user training related or have some other rational such as change deployment or concurrency locks ?

Stockland uses the Pulse’s user analytics feature to find all the long running operations where users have been waiting. It greatly helps to analyse the system behaviour from a performance perspective.

Using the Pulse’s user analytics feature, the admin team was able to identify a recurring long running thread incurring wait time for the user and the concurrent activities occurring at the same time.

The team was then able to take appropriate measures in agreement with the users to resolve the contention.

With Pulse, the admin team was able to go back to the user and explain to him why he was waiting and propose a plan to resolve the contention. It was identified that the user was scheduling a report update concurrently to scheduled chores targeting the same objects late in the evenings. It was then decided to re-sequence both the report update and the daily chore execution times.

One specific user was complaining that TM1 report updates were very slow to run. The TM1 administrators was able to find out in Pulse all the operations which were taking more than 10 seconds and narrow down to the contention. They found out that the user was trying to automatically update the TM1 reports at 10 pm while re-sequenced nightly chores had been set to run. Pulse allowed the TM1 administrator to investigate the issue and come up with a resolution by rescheduling the appropriate chore and determining the best timing for the report update to occur.