I am an application/backend developer who has to quibble with databases more often than desired. I can get my way around Postgres pretty well, but I can always use a hint or two, especially when it comes to fine-tuning and performance.
I stumbled upon Cleaning Up Your Postgres Databases. It offers useful advice on spotting performance bottlenecks in your Postgres database. Take the cache and index hit queries, for example.
The first thing you’re going to want to look at is your cache hit ratio and index hit ratio. Your cache hit ratio is going to give the percentage of time your data is served from within memory vs. having to go to disk. Generally serving data from memory vs. disk is going to orders of magnitude faster, thus the more you can serve from memory the better. For a typical web application making a lot of short requests I’m going to target > 99% here.
I will be trying them real soon. Like, today.