Service Level Objectives
Set stability targets for your Floods.
A Service Level Objective, also known as an SLO, is a set of targets for one or more measurements that you want your service to meet, under load, to be considered successful. Depending on your use case, the service you are testing may have hard requirements for success. SLOs are a way to encode those requirements as part of the load test itself.
This Flood passed its SLO for Mean Concurrent Users.
As an example, imagine a service which has a requirement to always keep response time below 2 seconds. SLOs let you define that contract as part of the test when creating, or editing, the Stream. Then, if any Flood launched from that stream broke the 2 second response time barrier the SLO would fail for that test.
Some SLOs will be more complicated than that, maybe requiring multiple measurements, targets and aggregations to determine failure. Each Stream can have as many SLOs as you'd like, and each SLO can be made from more than one rule.
A Stream with multiple SLOs defined.
We currently support the following measurements:
  • Concurrency
  • Failed Transaction Rate
  • Passed Transaction Rate
  • Response Time
  • Throughput
  • Transaction Rate
Aggregations:
  • Mean
  • Median
  • Min
  • Max
  • 90th Percentile
And for each rule, you can check if the value is >, , <, , = or to your target.
We are always looking to improve the measurement, aggregations we make available to you and welcome any suggestions for features you are missing.
Last modified 7mo ago
Copy link