Symptoms: a save, graduate, or load operation stays "in progress" far longer than usual, or eventually fails with a timeout error.

Likely causes:
* The underlying storage cluster is under load or degraded — this is more likely on smaller (e.g. single-node/embedded) clusters, where storage has less headroom to absorb concurrent operations.
* A previous operation on the same data container was interrupted (e.g. cancelled, or failed partway through) and left behind storage artifacts that are still being cleaned up.

What to do:
# Retry the operation — transient storage load often clears on its own.
# If it keeps failing on the same data container, check whether an earlier operation on it was cancelled or failed; wait a few minutes for background cleanup, then retry.
# If the problem persists, raise a support ticket and include a support bundle so we can inspect cluster storage health.