Analisar de onde vem o I/O com pg_stat_io (tipo de backend, objeto, contexto)

Comando

SELECT backend_type,
       object,
       context,
       reads,
       pg_size_pretty(reads * op_bytes) AS read_size,
       writes,
       pg_size_pretty(writes * op_bytes) AS write_size,
       extends,
       hits,
       round(100.0 * hits / NULLIF(hits + reads, 0), 1) AS hit_pct,
       evictions,
       reuses,
       fsyncs,
       round(read_time::numeric, 1) AS read_ms,
       round(write_time::numeric, 1) AS write_ms
FROM pg_stat_io
WHERE coalesce(reads, 0) + coalesce(writes, 0) + coalesce(extends, 0) + coalesce(evictions, 0) > 0
ORDER BY coalesce(reads, 0) + coalesce(writes, 0) DESC;

-- Who writes? Ideally most writes come from the checkpointer and background writer
SELECT backend_type,
       pg_size_pretty(sum(writes * op_bytes)) AS written,
       round(100.0 * sum(writes) / NULLIF(sum(sum(writes)) OVER (), 0), 1) AS pct_of_writes,
       sum(fsyncs) AS fsyncs
FROM pg_stat_io
WHERE object = 'relation' AND writes > 0
GROUP BY backend_type
ORDER BY sum(writes) DESC;

O que faz

Mostra leituras, gravações, extends, evictions e fsyncs por tipo de backend (client backend, autovacuum, checkpointer...), objeto (relation/temp relation) e contexto (normal, vacuum, bulkread, bulkwrite); e de quais tipos de processo vêm as gravações.

Quando usar

Durante um gargalo de I/O, para saber se a carga vem das consultas, do autovacuum ou dos checkpoints e para validar as configurações de shared_buffers/bgwriter.

Atenção

pg_stat_io é de todo o cluster e cumulativo desde o último reset; tire a diferença entre duas amostras para ver a variação (para zerar: SELECT pg_stat_reset_shared('io');). read_time/write_time só são preenchidos com track_io_timing = on. reads também inclui leituras atendidas pelo page cache do SO. Muitas gravações ou evictions de client backend significam que os backends gravam eles mesmos os buffers para liberar espaço; revise shared_buffers, bgwriter_lru_maxpages e as configurações de checkpoint. Linhas com context = 'vacuum'/'bulkread' refletem o uso de ring buffer; reuses ali são normais.