See CHANGELOG. It is a breaking change for read-only users, but IMO necessary, as timeout issues when an LB terminates a long-running query connection are incredibly obscure, and this should at least cover AWS LB defaults.
It works similarly to how clickhouse-connect handles it (but currently does not normalize it based on the request timeout).
Checklist
[x] Unit and integration tests covering the common scenarios were added
[x] A human-readable description of the changes was provided to include in CHANGELOG
[x] Docs will be updated if we decide to merge this.
Summary
See CHANGELOG. It is a breaking change for read-only users, but IMO necessary, as timeout issues when an LB terminates a long-running query connection are incredibly obscure, and this should at least cover AWS LB defaults.
It works similarly to how clickhouse-connect handles it (but currently does not normalize it based on the request timeout).
Checklist