Limit concurrent logic app (standard) workflow instances

Question

I have a scenario where I receive messages from service bus to trigger a workflow. This workflow does some processing but ultimately inserts some data into SQL DB. When 100,000's of messages appear at once, the DB gets overwhelmed.

Is there a way of restricting the number of concurrent instances?

Here is my test

This receives a non session message, has a 2 minute delay, then a compose just to ad an end activity.

If I submit 10 messages, enable the workflow, immediately all 10 messages cause 10 workflows to activate.

Here is my host.json file

Answer 1

Logic App standard runs on Azure Function so you should be able to limit the numbers of concurrent calls using host.json settings:

{
    "version": "2.0",
    "extensions": {
        "serviceBus": {
            "clientRetryOptions":{
                "mode": "exponential",
                "tryTimeout": "00:01:00",
                "delay": "00:00:00.80",
                "maxDelay": "00:01:00",
                "maxRetries": 3
            },
            "prefetchCount": 0,
            "transportType": "amqpWebSockets",
            "webProxy": "https://proxyserver:8080",
            "autoCompleteMessages": true,
            "maxAutoLockRenewalDuration": "00:05:00",
            "maxConcurrentCalls": 16,
            "maxConcurrentSessions": 8,
            "maxMessageBatchSize": 1000,
            "sessionIdleTimeout": "00:01:00",
            "enableCrossEntityTransactions": false
        }
    }
}

You could also set this setting using appsettings (see existing answer):

AzureFunctionsJobHost__extensions__serviceBus__maxConcurrentCalls

Answer 2

You can apply concurrency control at the app level (similar concept to what Thomas suggested, but there are specific toggles for Logic Apps). The link below points to those values:

https://learn.microsoft.com/en-us/azure/logic-apps/edit-app-settings-host-settings?tabs=visual-studio-code#trigger-concurrency

The value you are looking is Runtime.Trigger.MaximumRunConcurrency (and this is applied in the hosts.json). There is some guidance in the document on how to setup host.json (scroll up a bit).

Another option is to remove the split on, so you receive an array with all the messages and use a for-each loop, where you have more control on loop concurrency.

I hope this helps.

Cheers, Wagner.