7k requests a minute to azure blob storage

**Describe the bug**
I have a 4-node setup that utilizes Azure ServiceBus event hub, AzureSynchronization, and CosmosDbPersistance. Under minimal load (eg no active workflows) I'm getting a constant throughput of 7,000 requests to azure blob storage, consuming about 8mb per minute. Specifically it's calls to GetMessageRead that's happening.

**To Reproduce**
Create a workflow core service in Kubernetes with multiple nodes, utilizing all features of the azure extensions.

**Expected behavior**
Less than 7k transactions a minute on an idle workflow instance.

**Additional context**
It will bump up to 10k requests when busy, but only ever drops back down to 7k a minute.

Reproduce the four-node Kubernetes setup using Azure ServiceBus event hub, AzureSynchronization, and CosmosDbPersistance, then measure idle Azure Blob Storage traffic. Start by tracing the GetMessageRead calls and compare idle versus busy behavior. Done means an idle workflow instance stays below 7,000 transactions per minute without breaking active processing.

Written by the indexing model from the issue text.