Slow things down to make them go faster
summary
Covers how PostgreSQL’s multi-process internals, ACID/MVCC behavior, and locking/latch interactions cause surprising throughput drops under very high concurrency or transaction rates. Explains measurable pitfalls (lock waits, contention) and concrete mitigations such as tuning queries and connection parameters, using connection pooling and replication, and adapting workload patterns.
AI-written summary · about
related talks
45mTransactions and Concurrency Control Patterns
48mUnderstanding and Reducing Application Contention
47mAnatomy of concurrent data access in PostgreSQL
50mExplaining the PG concurrency control mechanisms using rollercoasters
41mConcurrency, Scalability and Transactions -- Myths and Surprises
48mModern SQL features in PostgreSQL
52mThe Long Transaction — When Size Matters
41mTransactions: myths, surprises and opportunities