Applying SLA to Guide Performance Testing and Boost Efficiency

The author, an Alibaba technical expert, introduces the sixth part of the "Performance Test Together" series, which aims to build a complete theory and practice framework for performance testing.

The article defines SLA (Service Level Agreement) and its components SLI (Service Level Indicator) and SLO (Service Level Objective), and outlines best practices for setting SLOs, such as specifying time windows and including exemption clauses.

SLA is divided into two dimensions: the business dimension (e.g., response time, error rate) that directly affects user experience, and the service dimension (e.g., CPU, load) that helps developers and testers locate problems.

In performance testing, a clear "performance test pass criteria"—derived from business expectations and engineering goals—acts as the testing SLA, linking external product SLA with more detailed internal metrics.

Using SLA during testing helps coordinate teams, focus on critical alerts, and decide when to stop a test, reducing manual monitoring effort.

The PTS platform supports SLA by allowing users to select metrics, set thresholds, and automatically trigger alerts or test termination when thresholds are breached.

Key steps for using SLA in PTS are: (1) understand the full set of client‑side (RPS, failure RPS, response time) and server‑side (CPU, memory, load, SLB, ECS, RDS) metrics; (2) choose core metrics and define appropriate thresholds; (3) run the test and let SLA alerts control test flow.

When SLA thresholds are hit, the test stops and a capacity report is generated, making the testing process more efficient and less error‑prone.

The article concludes that proper SLA definition and usage eliminate chaos during performance testing and pave the way for future intelligent testing integrations.