Concept of a causationId / "commit sequence" might be useful?

[PeterStephenson]

In the recent issue we found it really useful to see which events were saved in the same transaction as it showed that there was an issue cross thread. Without this information diagnosing the issue would have been very difficult. I can imagine situations where this is similar (especially where events can written out of order and can be triggered by multiple handlers)

Is it worth adding something like this? It doesn't need to be a wrapper or anything possibly just a guid indicating what was saved at the same time?

[DaveAurionix]

We added a correlation id to the event metadata, I don't know if you could do something similar? Correlation Id won't be unique to a commit, but you could literally add a CommitId.

[asosMikeGore]

Using the metadata for this is a good fit and doesn't require a change. It also means that you can then add additional items e.g. a causation id as well as correlation.

If this was added to the schema then we'd want the clients to master this, so they can use any identifier they need. A "transaction id" by itself (i.e. we master the id) likely isn't too helpful if you think about correlating this id across different activities across multiple services.

Is it worth us adding this in based on this, or just use the metadata? My preference right now is the latter.