[Change] Explain per-chunk nature of autovacuum

[ClaytonJY]

Change in content, appearance, or functionality? Addition of content

Describe the change I think it would be helpful to add something about how autovacuum works with hypertables: if you look up last_autovacuum/last_autoanalyze in pg_stat_user_tables for a Timescale hypertable, they are always null. This is because the autovacuum daemon operates on a per-chunk basis, so those fields are populated for chunks (schemaname = '_timescaledb_internal').

This is quite minor, but it may save someone else some digging in the future!

Location(s) unsure

How soon is this needed? No rush whatsoever!

[jonatas]

Very good suggestion @ClaytonJY! I'll make sure we consider it in our documentation roadmap.