Menu

Find Slow Database Queries

Requests and background jobs can be slow for various reasons, but a common cause are slow database queries. Large tables and missing indexes may cause database queries to take longer than needed, increasing the amount of time that your application's users have to wait.

AppSignal's Slow Queries feature makes it easier to find an application's slowest queries, to help you decide which queries require optimization.

You can view your application's slow queries here.

Finding Slow Queries for Ruby, Elixir, Node.js and Python

The type of events that can be found in the Slow Queries page are listed below.

  • Elasticsearch
  • MongoDB
  • MySQL
  • PostgreSQL
  • Redis
  • SQLite

Events may be grouped under the ORM (Object-relational mapping) event group name for some integrations.

Finding Slow Queries for OpenTelemetry

For OpenTelemetry-based integrations, we look for spans with a certain set of attributes listed below.

For the query body, we look for the following attributes in order:

  • db.query.text
  • db.query.statement
  • db.stored_procedure.name

One of these attributes must be present, and not empty.

For the system name, we look for the following attributes in order:

  • db.system.name
  • db.system

If neither is present, the system will be shown as Unknown.

If your application is reporting queries in a differently named attribute, or you would like to customize the recognized attributes, please give us feedback.

Sorting Queries

By default the slow query list is sorted by impact: the queries that take up the most time either by duration of queries or by how often they are performed.

It is also possible to sort this table by "Throughput", how often the query has been recorded, and "Mean", the mean duration of the queries.

Query Details

Query details screenshot

You can click on a query to gain the following insights:

  • Full query with optional event title for more context
  • The database query event summary:
    • Mean duration of query
    • Throughput, how often the query was recorded
    • The impact of the query in percentage of all database queries in the selected timeframe.
  • Response time graph
  • Throughput graph
  • Actions and performance measurements in which the event occurred:
    • You can go directly to actions and performance measurements in which the event was recorded to view the impact per request/background job in the sample event timeline.