I am looking for the details on the checkpoint mechanism in the Azure Logic App when the “When events are available in Event Hub” trigger is selected.
My simple setup as shown below:
Details:
- I have an Azure Event Hub namespace with one Event Hub in it. The partition size is 4. There’s a consumer group called “MyConsumerGroup” on this topic.
- For testing, using the local Python code, I publish simple events to this topic.
- I also have an Azure Logic App; the setup is as shown. This listens to the “MyConsumerGroup”.
- As soon as the events are published, the logic app is triggered with all the default settings and writes the events to the Blob Storage.
The happy flow of this runs smoothly.
I am looking for the negative case wherein e.g., my storage container is deleted, the trigger would be invoked but it won’t find the container. So the logic app workflow would fail as shown.
Now, when the container is back again, ideally it should process these 3 event first in their sequence and then process any new events as per the standard checkpoint strategy.
I tried to replicate this but it processes only the new events, entirely skipping these 3.
After going through blogs and videos, I found out that the Logic App maintains the checkpoint internally and it is automatic.
I want the Logic App workflow to start from the last processed event, then read all those which are failed first and then proceed with the new events.
Somehow, I am having hard time to validate this.
Questions:
- Can we modify the checkpoint? If yes, where’s that option?
- Is there any other alternative to test this or make it work?
- Is it something about the 4 partitions and that the events are not in sync?
Can someone please help?
Thank you!
Sanket