Query Analyzer
Load-intensive queries are a source of slowdowns and can increase time to diagnose a problem when one occurs. Queries that are called often, or that have a high load can lead to slow response.
Query Analyzer allows users to understand the relative load of the automated queries like monitors and recording rules, ordered from most to least problematic. Using this information, users can assess whether they need to make modifications to the queries. Query Analyzer uses data from the previous day (24 hours UTC) to generate the report.
Queries with a high load can potentially trigger query protections.
Use the analyzer
In the navigation menu, click Go to Admin and then select Analyzers > Query Analyzer.
Each entry is a single line. Click any line to expand that entry to display the entire query.
The table displays the following data:
- Query Reference: The name of the recording rule or monitor. For monitors, click the monitor name to go directly to the monitor page.
- % Load: (sortable, default) The percentage of overall query load this query contributes.
- Times Called: (sortable) The number of times this query was called during the time range.
- Query: The query the monitor or recording rule uses.
The total number of displayed rows follows the chart.
Find specific data
Use the Search box to find exact matches for text in the Query reference or Query fields.
For example, searching for k8s searches all references and their queries for that
string. While reference names might not include k8s, many queried metrics do. A
recording rule named server-availability-error-ratio-4w could include
grpc_server_handled_total:without_instance{_k8s_namespace="global"} in the
query, matching a search for k8s.
When you search for specific data, the display updates:
- The table displays all matching items.
- The number of rows shows the number of rows that match the search.
- Summary cards for Matching Queries, % Load, and Times Called display. These cards are a sum total of the rows matching the search criteria.