Last updated 31 August 2018
Expensive queries are database queries that run slowly and/or spend a significant amount of their execution time reading and writing to disk. Such queries are the most common cause of performance issues on Heroku Postgres databases.
Optimizing expensive queries can significantly improve your application’s performance and overall response times.
Viewing expensive queries
You can view expensive queries for your database at data.heroku.com. Select a database from the list and navigate to its Diagnose tab.
The Diagnose tab is not available for Hobby tier databases.
Queries for each category (most time consuming, most frequently invoked, etc.) are shown, with the most likely candidates for optimization at the top. Each query has accompanying graphs that show how often the query executes and how much time it usually takes. Expensive query data is available for up to the last seven days.
Identifying slow parameter values
If a slow query uses bind parameters (
$1) or query parameters (
?), the Diagnose tab does not show the values of those parameters. Performance can vary significantly between different parameter values for a given query.
To help you identify particularly slow parameter values, Heroku outputs the slowest queries (that take 2 seconds or more) and their parameters to your application’s logs if you enable slow query logs. You can use a logging add-on to search your log data for these slow queries.
Jun 26 08:49:40 issuetriage app/postgres.29339: [DATABASE] [41-1] LOG: duration: 2406.615 ms execute <unnamed>: SELECT "issues".* FROM "issues" WHERE "issues"."repo_id" = $1 AND "issues"."state" = $2 ORDER BY created_at DESC LIMIT $3 OFFSET $4 Jun 26 08:49:40 issuetriage app/postgres.29339: [DATABASE] [41-2] DETAIL: parameters: $1 = '1348', $2 = 'open', $3 = '20', $4 = '760'
This query took 2.4 seconds to execute. Its parameter values were
20 for the query’s
760 for the query’s
OFFSET. You can use this information with
EXPLAIN ANALYZE to profile why a particular parameter combination is slow.
Causes of expensive queries
The most common causes of expensive queries are:
- A lack of relevant indexes, causing slow lookups on large tables
- Unused indexes, causing slow
- An inefficient schema leading to bad queries
- Inefficiently designed queries
- Large database size or high lock contention, causing slow
COPYoperations (usually used for logical backups).
Solutions to expensive queries
Here are some guidelines that might help fix expensive queries:
EXPLAIN ANALYZE(via pg:psql) to find out what’s taking most of the query’s execution time.
- For example, a sequential scan on a large table is often a bad sign. Efficient indexes can improve query performance dramatically. Consider all Postgres techniques, such as partial indexes, when devising your index strategy.
- Identify unused indexes by running
heroku pg:diagnose. Drop any indexes that are not required.
- Upgrade your database to the latest version. Postgres performance improves with virtually every release.
- For large databases, prefer relying on Heroku’s continuous protection for day-to-day disaster recovery purposes. Remove any auto
pg:backupsplans, and use
pg:backupsstrictly for extracting or migrating data.
- For smaller databases, slow logical backups can be a result of lock contention.
pg:outliersfrom Heroku pg-extras to find queries that have a high proportion of execution time.