Pattern for CUD operation

In many enterprise applications, implementing a Create/Update/Delete operation requires performing the following actions:

• Updating one or more pieces of information in a database
• Communicating with one or more external systems
• Refreshing/invalidating some in-memory cache

My question is: What is the best way to implement the steps above so that there is consistency/eventual consistency across DB, external systems and cache in the face of computer or network failures?

I understand that, if DB + external system + cache are 2-phase commit capable, the whole thing can be kept consistent using a global transaction. But that's rarely an option.

One approach I can think of is as follows:

1. In a single DB transaction, perform the updates and create a new "workflow" object with status=PENDING and containing enough information to communicate with the external system and refresh/invalidate the cache.
2. Communicate with the external systems, in a idempotent manner
3. Refresh/invalidate the cache
4. Set the status of the workflow object to COMPLETE (or just delete it)
5. Have a periodic background task that searches for workflow objects with status=PENDING and that have been sitting idle for too long. For each of those objects, repeat steps #2, 3 and 4 above.

Does the approach above seem reasonable? Are there better alternatives?

Thank you in advance for any comments/suggestions.