Camunda batch process scaling

Camunda 7 Topics Discussion & Questions
1

Camunda batch process scaling

2

Team ,

I am looking for a approach to scale custom batch process in terms cluster node.

Lets say I have a batch Jobs which runs 1million accounts and each of account number get with some process definition flow (process instance) .
Now as per environment is concern I have kubernetes cluster where I have 2 PODS or more PODS running . So question is how the load of 1 millions accounts share among the PODS ,even I have custom parameters like batch size , Seed size

I did some analysis on my local .
1:Runs one batch job , which execute N no of account process instance batch by batch as per custom parameter .
2:As its springboot application , I run the application into 2 port (6001/6002) .
3: I am able to load the cockpit for both port in 2 browser .
4: Now I create account process instance form both port like below
http://localhost:6001/myService/engine-rest/process-defination/key/AccountNumberRecevied/start
http://localhost:6002/myService/engine-rest/process-defination/key/AccountNumberRecevied/start
5:So what I observed for executing the batch its always take one port either 6001 or 6002 , its wont share’s the load among each others.
6:Means camunda batch is works on singleton frame work .

Finally , when I run the process instance form both port and stop one of the instance , then other port look all the load and execute the batch .

So is there any better way in camunda batch which help scaling the load among PODS in kubernetes cluster environment ?

Thanks

3

We need to tune in order to work on batch processing. I would recommend to read the configuration part.

https://docs.camunda.org/manual/7.16/user-guide/process-engine/batch/#configuration

4

Thanks for your answer .

Above link its specify the implementation part base on batch configuration and i am good with that .

Question is mainly on load balancing on among the PODS , if one POD is busy on executing Job as per configuration parameter then how its share the load to other PODs ? Mostly question is how to scale PODS based on load ?

5