The ultimate Fabric, Power BI, SQL, and AI community-led learning event. Save €200 with code FABCOMM.
Get registeredCompete to become Power BI Data Viz World Champion! First round ends August 18th. Get started.
After facing some performance bottlenecks I started looking into the Server Timings page in Dax Studio to better understand what's going on.
As you can see this expression took more than 45 seconds. We then look at the individual rows and try to se which ones take the most time and try to optimize that part.
Question 1)
There is a duration column, but summarizing those values only gives us the SE queries duration?
Question 2)
If we look closer at rows 86 and forwards we see there is a huge leap from the timings indicator even though the last operation only took 1 ms. What's going on here? Is it the Formula Engine cranking for dear life without any traces? How would i go about to optimize this..?
Question 3)
Do you recognize this charcteristic of the timeline that we have two huge operations more or less and can explain in general what could be poorly configured/usual suspects...?
Solved! Go to Solution.
We are trying! Currently we have identified odd performance behaviour with
SWITCH
Selectedvalue(Parameter1),
"Option1",Measure1
"Option2",Measure2
"Option3", DefaultMeasureIfNothingSelected (but as an option...)
DefaultMeasureIfNothingSelected
Where the default of the switch is 2 seconds (6->4) faster than if we have selected the option that will return the same measure.
you have 50 queries that are 99.9% using the formula engine. You need to change the way you wrote your formula. Try different variations. Do not forget to use variables and optimize the number of queries as max as possible
We are trying! Currently we have identified odd performance behaviour with
SWITCH
Selectedvalue(Parameter1),
"Option1",Measure1
"Option2",Measure2
"Option3", DefaultMeasureIfNothingSelected (but as an option...)
DefaultMeasureIfNothingSelected
Where the default of the switch is 2 seconds (6->4) faster than if we have selected the option that will return the same measure.
User | Count |
---|---|
26 | |
10 | |
8 | |
6 | |
6 |
User | Count |
---|---|
31 | |
11 | |
10 | |
10 | |
9 |