If you’d like, you can construct an issue for it in the issue tracker. I thought about this a little, and I think we can introduce some helpful solutions on the subject.
Yes, I can, although I think it has to do with machine resources (massive hunch). It’s not overly clear why the connection between your Axon Framework application and the Axon Server instance is closed. More specifically, why it became UNAVAILABLE.
The only guess I’d dare to make is the memory issue you’ve also shared.
Anyhow, I do not see how these traces lead to the deserialization problem we started with. It might be a stretch, but could this scenario somehow relate to this issue you’ve opened? Honestly, snapshots of that size are not regular by my standards.
Although it happened in production, I am glad you took this route. Changing the revision number on the aggregate triggered Axon’s RevisionSnapshotFilter to disregard snapshots that didn’t match the current revision. As such, you basically filtered out the “unserializable-snapshot.”
Concluding (for now), my sample project request was for the serialization issue.
Not for the connection issue, you might’ve experienced because of machine resources."
So, whether you’d still want to traverse that route is up to you. From my end, I’ll ask one of the Axon Server engineers if this situation rings a bell.