Which process id gets deployed in a "pool collaboration"

Hi Team,
I was learning the Camunda Academy topics and came across a concept of “pool collaboration” where we can model 2 or more pools and communicate with each other. Is this workable with Camunda 8?
Few quires on this:
Each pool corresponds to each process. So when we deploy 2 pools (aka 2 processes) I observe that the process ID of the latest pool implemented only gets displayed in the Operate dash board. Hope this is expected.
How exactly does the pools collaborate with each other in real time. I tried the below simple sample but unable to logically execute this.
image

Is there any sample that shows the pool collaboration.

Thanks,
Saju

Hi,

all pools that are defined as executable are deployed (click on the pool label then check the properties panel for the “Executable” checkbox under “Process name”).
At the moment there is a visual bug in operate that only shows one of the process names for all deployed processes.

Two or more executable processes in one model are basically independent.
Pools and message flows themselves are not executable, so the meaning of your model is: The user who works on task “Pool1 A” may manually start the process within Pool2.
If you want to ensure that the Pool2 process is started, you can use message events or message tasks (Pool1 sends a message, Pool2 reacts to it). Alternatively, without pools, you can use subprocesses or call activities.

Kind regards

Thanks @DatyEighty

Technically I see that having multiple pools can’t add value in operate as the processes are separately tracked. Eg. If I deploy the below bpmn
pool_collaboration.bpmn (10.9 KB)
and start a instance (from modeler) , only Pool1 gets invoked and if I invoke a new message start event it creates a new instance entry of Pool2 in operate.

Operate see’s the deployment with 2 pools as separate processes and tracks them separately.
So eventually the transactions might not be readable as we see in modeler when it comes to operate.
So when the above bug gets fixed can we expect both pool1 and pool2 transactions in operate also will be tracked together.

Thanks,
Saju

Hi @Saju_John_Sebastian1

I’d like to add that this is in line with the BPMN specification. Each pool represents a different process.
Having two pools with processes in one diagram is the same as having two processes in separate diagrams. These processes are not directly related. However, you can use other BPMN elements to make processes work together.

The operate bug is a visual thing only. It doesn’t affect how the model is interpreted.

The modeler also has some limitations. It doesn’t handle multiple processes in one diagram well (e.g. you can’t select which process to start). Also, it doesn’t allow you to start message events (yet).
If you want to make use of these features, I suggest using the command line interface zbctl.

Kind regards