What’s the max size for the Redis Trigger on an Edge Workflow?
It works great with small objects until I include a payload that’s about 300k in size. (an image encoded in base64)
Redis-cli subscribe shows the subscription event, but the Losant Redis Trigger never registers this event, nor shows any errors in the log, it’s just silent.
If I remove the larger payload object, the Losant Redis Trigger works fine again.
Your workflow should still be firing, but the payload may be too big for the Debug log entry, which has a limit of 256K.
To test this, can you try to put in a singular property from the payload other than the base64 encoded image in your Debug node similar to this:
This will show just a single property from the incoming payload rather than the whole payload and hopefully give an indication that the workflow is running as you expect.
Please let me know the results of this.
Oh, that is unfortunate, and matches what I saw. Short strings work fine in the same field.
As a suggestion, please consider truncating those larger values to create a consistent debug experience. Now I need to figure out how to engineer a new flow pattern that I can debug.
We do have an open feature request for this exact thing. I’ll be sure to give your +1 to this and keep you in the loop with what I learn.
I want to confirm, though, that the workflow you have is still firing even though you are unable to view the debug messages? I just want to verify that there aren’t any other issues.
@Heath - you were correct and so glad to hear about the +1 feature request!
I modified my debug to only display data.topic, so I perhaps could confirm if the workflow was running or not by just confirming the topic received.
I also added a HTTP step that would post into a chat room if the workflow ran.
What I found is what you described. Redis-cli shows the publish on the channel, and the workflow fired as you will see when you read on, but when the payload is over 256k, the debug step - even limited to just display the topic, would not fire.
The HTTP step fired though, even after failing to display the debug data.
So yes - the workflow runs, but there is no tooling to debug any part of it, which would be very challenging to support.