Timesheet Performance lag in calculating total hours per day in the timesheet period of each cell
search cancel

Timesheet Performance lag in calculating total hours per day in the timesheet period of each cell

book

Article ID: 397312

calendar_today

Updated On:

Products

Clarity PPM SaaS Clarity PPM On Premise

Issue/Introduction

On a high load timesheeting day, the Total Hours per day in a time period calculation is slow and this is also causing the DB CPU to hit 100%.


STEPS TO REPRODUCE
1. Identify a Clarity machine having large Timesheet dataset, typically a 10k+ user base and with atleast 5+ yrs of timesheeting history.  
2. Make sure the below conditions are met 
   --> 50+ timesheet users entering timesheets at the same time. 
   --> atleast 3 timesheet related processes in Active state
3. Login as a Timesheet user and navigate to a timeperiod (MUX)
4. Enter 8 hours each from Mon - Fri
5. Check how quickly the Totals per day is calculated and reflected in the per day Total column

Expected: The Totals per day is reflected immediately (less than a second)
Actual: Depending on the load on the system, it takes more than several seconds to reflect the hours from a single cell

Workaround: None

Environment

16.3.1

Resolution

DE170431 - Fixed in 16.3.3, 16.3.2.1

Note:  When the toggle is enabled all time entry update events are suppressed. As a result if there is a process start condition on time entry update, the process will not auto trigger.

Please raise a Support case if you experience this issue and need to enable the toggle. 

Additional Information

--> The DB CPU hits 99-100% fairly quickly when there are 50+ users trying to enter timesheets. Submitting the timesheet is not required. 
--> Each enter in the cell is making a REST API call
--> The below query which should run farily quickly starts taking almost a second
   SELECT s0.prid odf_pk, s0.prresourceid AS prresourceid FROMprtimesheet s0 WHERE (s0.prid = ?)