Loading...

XML

Word

Printable

given:
- I have an engine 7.8.0 with a postgres database running.
- the engine contains a huge amount of data (e.g. > 2 million process instances, > 20 million activity instances, > 20 million variable instances)
when:
- I perform a count query on the historic activity instances or historic process instances or historic variable instances
then:
- the request takes a very long time, such that it times out (e.g. > 20 seconds).
expected:
- the request takes at max 2 seconds

Hint:

Checkout the following engine rest endpoint with a huge data set to reproduce the behavior (use your LDAP credentials):
https://performance.optimize.ci.camunda.cloud/engine-rest/
Postgres is known to have problems with the table count. We should check if postgres is doing a full table scan every time we query the count.
Also it might be worth checking if other databases cause similar problems.

is related to

CAM-10282 Weak performance of count queries on PostgreSQL