EventHub partitions are not equally spread across all service instances;

[dmoeff]

I have Kubernetes service with 32 instances connected to Azure EventHub with 32 partitions. Some of the instances of the service are not able to connect to any partition, where other are connected to 2 partitions. My expectation is every instance of the service to process events from a single partition when the number of instances is equal to the number of partitions.

My service is using github.com/Azure/azure-sdk-for-go/sdk/messaging/azeventhubs v1.0.2 package and is running on AKS cluster. It is built on top of ConsumingEventsWithCheckpoints example using go version 1.21

[richardpark-msft]

How long does this unbalanced situation last? Are we talking "never resolves" or "resolves after a few minutes?"

[richardpark-msft]

@dmoeff, I was able to reproduce this with a better test than I had. I have the draft PR out, and will get it out early next year.

[ilyasse05]

i have the same problem look forward to the solution ;)