Version 2021
Tuning
The following section describes tuning PostgreSQL for analytical workload based on internal tests at MicroStrategy.
|
Parameter Operation |
Initial Value |
Changed To |
Performance Gain |
% Execution Time Of Original Value |
|---|---|---|---|---|
|
shared_buffer |
128 MB |
1280 MB |
Metadata garbage collection time reduces from 5724 seconds to 47 seconds. |
0.82% |
|
work_mem |
4 MB |
50 MB |
Metadata garbage collection time reduces from 47 seconds to 17 seconds. |
36.2% |
|
maintenance_work_mem
|
64MB |
1536MB |
Four times performance gain on the index creation on platform_analytics_wh.access_transactions table. |
25% |
|
max_parallel_maintenance_workers |
2 |
8 |
||
|
random_page_cost |
4 |
1 |
Execution time of query “select tran_id from access_transactions as a join temp_lu_object as b on a.object_id=b.object_id limit 1000;” in Platform Analytics is reduced from 214185 ms to 2.3 ms. |
<0.01% |
|
enable_nestloop |
on |
off |
Table scan time reduced from 585 seconds to 8 seconds. |
1.36% |
|
vacuum |
NA |
NA |
Table size reduced from 110 GB to 1.5 MB after vacuuming the table platform_analytics_wh.access_transactions |
<0.01% |
See KB485086: Best practices for performance tuning based on PostgreSQL for details.
ODBC Versions 2 and 3
Starting in MicroStrategy 2021 Update 1, MicroStrategy uses ODBC call 3.x for PostgreSQL by default.
For a single cube publish, using 3.x ODBC calls produced an approximate 30% performance improvement over using 2.0 ODBC calls.
The test shown below is based on a four million row dataset which shows the throughput improvement with ODBC call 3.5 compared to ODBC call 2.0.

Data Engine Version Upgrade Since MicroStrategy 2021
During MicroStrategy performance testing with a data engine version of 2021 using MicroStrategy 2021, there was an overall 35% performance gain compared to data engine version 10.4.
Starting in MicroStrategy 2021, the default data engine version is 2021. See KB485086: Data Engine changes in the MicroStrategy 2021 release for more information.
For example, with a data engine version of 2020 and above, MicroStrategy pushes filters down to the database for execution to minimize the intermediate result set size. This pushdown can significantly improve the performance of dossier execution. For more details, see KB483573: Metric Qualification Filter Not Pushed Down in ConnectLive Leads to Performance Degradation.
The following chart shows the throughput change with different data engine versions. The throughput of data engine version 2021 is higher than data engine version 10.4. The test was executed with 50 concurrent connections.
