Camunda Cluster

Hi,
I would like to scale our single tomcat instance running camunda, cockpit and the camunda rest-api. What I would like to know is, what additional changes would be required to allow the multiple instances represent a single cluster (2 separate nodes running their own tomcat and camunda instances). I have currently configured 2 instances of tomcat + camunda and setup session affinity via nginx to allow access to the cockpit dashboard. Both instance’s has got their bpm-platform.xml configured to StandaloneProcessEngineConfiguration and are pointing to the same database. Ideally I would like the 2 instances balance/share the camunda workflows between them. When I currently do a GET call to the REST API /engine path I only receive back one entry (default). I would expect to get 2 (default) entries for the processing engine but this is not the case. Could someone perhaps guide me in a direction as to which additional changes are required to have the 2 instance work together as one cluster?

1 Like

Hi,

The behaviour you see is correct. Think of an engine as a logical construct rather than physical node. Hence, whilst you have two processing nodes, you still have one logical engine.

The two instances should be working as a cluster. Given they are using the same DB, the job executors will share the background load. Foreground requests will be shared according to your loadbalancer configuration…

This whitepaper make also give you some insight…

regards

Rob

1 Like

Thank you for your quick response. This answers my question.