Get the latest tech news

PostgreSQL Statistics, Indexes, and Pareto Data Distributions

Growth, Statistics, Indexes, And Keeping Our Dialer Humming A Brief Introduction To Our Dialer System We handle a fairly large amount of calls at Close, helping our customers connect with their customers ~9 million times a month. The core service that we’ve built to manage the lifecycle of each call is called the Dialer.

This command makes it so that, instead of scanning a sample of 30,000 rows (the default value of 100 multiplied by a magical constant of 300 present here and ), our PostgreSQL will look at 3,000,000 Dialer calls to gather the updated statistics. – almost nobody makes calls at that time, so even with the increased statistics target, PG found approximately zero rows with null date_terminated, and reverted back to using the suboptimal index. You can either partition the table or – if you don’t need the data past some expiration date – you can set up automatic purging of the old entries (leaving enough buffer for forensics).

Get the Android app