Skip to main content

New Relic Queries

This document contains a useful query for New Relic to count transactions within a specified time frame and filter by application name.

Query for Transaction Count​

SELECT count(*) FROM Transaction WHERE appName IN ('service-production') AND name LIKE '%queryName%' SINCE '2024-08-31T20:00 [America/Bogota]' UNTIL '2024-08-31T20:15  [America/Bogota]' FACET request.headers.origin

Explanation​

This query counts the number of transactions for a specific application within a defined time range. Here’s a breakdown of the command:

  • SELECT count(*): Selects the count of all transactions that meet the specified criteria.

  • FROM Transaction: Indicates the data source, which in this case is the Transaction event.

  • WHERE appName IN ('service-production'): Filters the results to include only those transactions where the application name is service-production.

  • AND name LIKE '%queryName%': Further filters the transactions to include only those whose names match the specified pattern (queryName).

  • SINCE '2024-08-31T20:00 [America/Bogota]': Defines the start time for the query.

  • UNTIL '2024-08-31T20:15 [America/Bogota]': Defines the end time for the query.

  • FACET request.headers.origin: Groups the results by the origin header of the requests, allowing for insights into where the requests originated.

Transaction Duration Metrics​

SELECT 
  min(duration) as min_time,
  average(duration) as average_time,
  percentile(duration, 50, 90, 95, 99),
  max(duration) as max_time,
  stddev(duration) as stddev_time
FROM Transaction 
WHERE appName = 'service-name'
SINCE 4 hours ago
-- SINCE 1 days ago
-- SINCE 30 minutes ago

Explanation​

This query calculates detailed performance statistics regarding transaction durations (latency). It provides a comprehensive view of how fast or slow the application is responding. Here’s a breakdown of the command:

  • min(duration) as min_time: Returns the shortest transaction duration found in the time window.
  • average(duration) as average_time: Calculates the mean response time.
  • percentile(duration, 50, 90, 95, 99): Displays the duration at various thresholds. For example, the 99th percentile indicates that 99% of requests were faster than this value (useful for spotting outliers).
  • max(duration) as max_time: Returns the longest transaction duration (the slowest request).
  • stddev(duration) as stddev_time: Calculates the standard deviation, which helps measure the consistency of the response times (a lower number means more consistent performance).
  • WHERE appName = 'service-name': Limits the analysis to a specific service.
  • SINCE 4 hours ago: Sets the time window for the analysis. The commented lines (--) show alternative common timeframes like 1 day or 30 minutes.

This query is essential for analyzing application health, identifying performance degradation, and validating Service Level Agreements (SLAs).