The number of blocks from indexes/tables that were found in the Postgres buffer cache.

This includes values for child operations and is a total value (i.e. not a per-loop average).

Once you have subtracted child operation values, the proportion of shared hit to shared read blocks is an interesting measure of cache performance.

The ratio of shared blocks (hit plus read) to rows processed (including any filtered out) is also interesting: a high ratio means Postgres is reading a lot of data for what it is doing, which can be a sign of bloat.

Related fields

• Shared Read

• Shared Dirtied

• Shared Written

• Shared I/O Read Time

• Shared I/O Write Time

Required parameters

• BUFFERS

Related official docs

• Resource Consumption

• Using EXPLAIN ANALYZE

• Database Page Layout

Further reading

• Issues with VACUUM and Bloat – Cybertec

• Using BUFFERS for query optimization – pgMustard

• Explaining buffers – depesz