Camunda Cluster Performance with Job Executer

Camunda 7 Topics Discussion & Questions
1

Hello guys,

I’m doing some tests with Camunda Cluster to optimize the performance of my executions.

The performance when I test a transactionally workflow is notable. The more servers, the faster it gets.

But now, I want to optimize the throughput using Job Executer.

My scenario is: If I use 1 embeded engine or 3 embeded engine using the shared database, my throughput at the end of the executions still the same.
I would like to have a better performance if I use more embeded engine.

I tried to set this properties at the embeded engine, without improvement in the result:
setCorePoolSize(20)
setMaxPoolSize(50)
setQueueSize(48)
setMaxJobsPerAcquisition(48)

Camunda version: 7.8.0
Database: PostgreSQL 9.6.6

Thanks!

Best regards,
Robert

2

Hi Robert,

  1. Do all those embedded engines share the same computation resources (i.e. machine)?
  2. What is your reasoning to choose these numbers?

Cheers,
Thorben