Hey Kyle, What version of Cassandra are you running? Can you share the full (anonymized) schema, including keyspace and table definitions?
The trace behavior you’re seeing here might be caused by read speculation: https://cassandra.apache.org/doc/5.0/cassandra/developing/cql/ddl.html#spec_retry Worth seeing if you can reproduce the issue with read speculation disabled for the table. Separately - I think there’s a bug in our tracing implementation I’ve noticed elsewhere, where it seems like part of the trace is cutoff and reset. I wouldn’t expect the coordinator’s first event to be issuing the data read to a replica (event_id fea78c60-3775-11f0-88c7-67d23f0ba311).
