Closed michaelkohn closed 2 months ago
contact_list:load
for potential improvementThe baseline "BEFORE" apdex score will be obtained from running the test suite in these three configs
To be confirmed
CHW
userFor CHWs
We will execute the automation tests for 1 user and simulate 5 days, with the suite running 10 times each day.
contact_list
According to the test plan documented in this comment. We expect to have 3 entries of contact_list:load
for each test run.
As an offline user, access the application and perform the following navigation:
contact_list:load
is recorded once)contact_list:query
is recorded once)contact_list:load
is recorded once)contact_list:query
is recorded once)contact_list:load
is recorded once)contact_list:query
is recorded once)contact_list:query
is recorded once)contact_list:query
is recorded once)At the end of the test (Running the test 10 times for 5 days) we expect to see:
150
entries of contact_list:load
250
entries of contact_list:query
Hi @michaelkohn
Hi @michaelkohn, due to the issues I am facing with the automated testing I decided to get a baseline using manual testing.
I ran the manual testing 30 times how I outlined it:
Apdex Score = ((Satisfied Count) + (Tolerating Count / 2)) / Total Count
Nice @Benmuiruri. Thanks for the clear data... it's really interesting to see how much worse Togo is, I wonder why??
I assume the Mali and Nepal users you set up are only seeing 2-3 rows in the contact_list (due to where they are in the hierarchy), so it's not surprising they are scoring very high. Since we have a good dataset of, well, bad data with Kenya and Togo, I'm fine if you just proceed comparing the results of before / after code changes for Kenya and Togo only since there really isn't any room to improve Mali/Nepal anyway. The one thing that could be useful, but not totally necessary, would be to show that Mali/Nepal didn't get slower after the code changes, but it's not necessary since I imagine it's a very low chance it gets slower.
contact_list