Post Content
Should you lower random_page_cost on SSDs in Postgres? Tomas Vondra (Microsoft) explores this question in his talk “random_page_cost in Postgres – why the default is 4.0 and should you lower it?” at POSETTE: An Event for Postgres 2026. Abstract: random_page_cost is one of the basic parameters affecting query planning in PostgreSQL. It expresses the cost of random I/O in relation to sequential reads. The lower the value, the “cheaper” the operations performing random I/O — typically Index Scans. Since 2002 the default value is set to 4.0, but how did we pick it and why didn’t it change over the ~25 years? The storage changed a lot over the years. And should you reduce it when running on SSDs?
Tomas Vondra (Microsoft) is a long-term PostgreSQL contributor & committer, focusing on performance, query planning & optimization, etc.
► Video chapters:
⏩ 00:00 – Music & introduction
⏩ 00:28 – Agenda and key questions to answer
⏩ 00:45 – How Postgres cost-based planning works
⏩ 02:05 – Sequential vs random I/O explained
⏩ 03:40 – Breaking down cost formulas and parameters
⏩ 07:50 – Where the default value 4.0 comes from
⏩ 10:40 – Re-running old experiments on modern hardware
⏩ 13:20 – Measuring real I/O costs with explain analyze
⏩ 15:35 – Surprising results from SSDs and spinning disks
⏩ 18:25 – Why higher costs can hurt overall performance
▶️ 22:37 – Should you lower random_page_cost?
▶️ 24:10 – Practical tuning advice and trade-offs
📕 Everything you need to know about POSETTE: An Event for Postgres can be found at: https://posetteconf.com
✅ Learn more: watch more POSETTE talks: https://aka.ms/posette-playlist
📌 Let’s connect:
LinkedIn: https://www.linkedin.com/company/posetteconf/
X – @PosetteConf, https://x.com/PosetteConf
Mastodon – @posetteconf, https://mastodon.social/@posetteconf
Bluesky – @posetteconf.com, https://aka.ms/posette-on-bluesky
#PosetteConf #PostgreSQL #database Read More Microsoft Developer
