As you may have noticed, this is still (somewhat) a P.o.C. There is no resiliency baked into this application yet, so don't throw anything mission critical on it yet. A non-exhaustive list of updates it would benefit from:
Server needs to be productionized and easy to deploy - Docker-Compose is recommended route.
Write up a more guides on how to self-host it.
Sending webhooks should have a simple retry mechanism in place, in case it just needs a few seconds before things work hunky dory.
Incoming actions should be placed into a resilient queue, that way events aren't lost in the case of downstream failure response, server outage, or etc.
A DB like SQlite can act as a queue in a pinch, so long as you setup an easy cron option. Rather than figuring out how to setup Celery with Python, why not use one of those services that will send you a webhook on a cron schedule, so all you have to write is an endpoint. Cronhooks is one such aptly named service.
Tracking the timestamps of when the workflows were kicked off and also whether they succeeded or failed would be a nice touch, but that's a lot of data for a PoC app.
As you may have noticed, this is still (somewhat) a P.o.C. There is no resiliency baked into this application yet, so don't throw anything mission critical on it yet. A non-exhaustive list of updates it would benefit from: