Developer
v2026.10
Updated 4 Nov 2026
Measure the workload
Check active sessions, long-running transactions, slow queries, table and index sizes, cache hit rate, locks, temporary file use, and disk I/O before changing instance size.
Common campaign pressures
Large dashboard queries, high-volume sync, wide exports, import jobs, and repeated geography joins can create different bottlenecks. Treat them separately.
Disk growth
Track database tables, indexes, application storage, report temp files, logs, and container image layers. Keep safe free disk and remove only artefacts covered by an approved retention rule.
Scaling decision
Optimise expensive queries and indexes first; then increase CPU/RAM or separate PostgreSQL when sustained load still exceeds safe headroom.
Was this page helpful?