feat(autogpt_builder): Ensure we have insight into flow sync conflicts (client <-> server)

[Pwuts]

It is undesirable but possible that the server — due to differences in logic or syncing issues — has a different version of a flow than the client. To circumvent this, we currently have the following workflow for saving/syncing a flow:

1. Save: send serialized flow to server
2. Retrieve stored flow from server
3. Replace local version of flow with retrieved version from server

In addition to this:

• we should be logging events where the server-side version (retrieved at step 3) differs from the version submitted at step 1.

This is essential for catching and patching data integrity issues.

We could log events like this to Sentry.