Setting up performance SLOs/SLAs and thresholds

Key Concepts

You can set project-level Service Level Objectives (SLOs), which can also be used to track Service Level Agreements (SLAs). When an API call sequence does not conform to the SLO thresholds you have defined, the system will trigger alerts; you'll also see violations in the reporting interface.

SLOs can be created based on the timing of the whole API call, or based on the timing for a single component of the call. You can set SLOs and receive alerts based on the following metrics (individually or in combination):

• Overall availability rate (as a percent of all API calls run)
• Overall pass rate (as a percent of all API calls run)
• Overall slow rate (as a percent of all API calls run)
• Overall warning rate (as a percent of all API calls run)
• Overall fail rate (as a percent of all API calls run)
• Total latency (in milliseconds)
• DNS lookup time (in milliseconds)
• TCP connect time (in milliseconds)
• SSL handshake time (in milliseconds)
• Total network time (in milliseconds)
• Upload time (in milliseconds)
• Processing time (in milliseconds)
• Download time (in milliseconds)
• Overall server time (in milliseconds)
• CASC quality score (based on score range)
• Apdex score (based on score range)

To set or change SLO criteria, navigate to Conformance > Conformance Settings, then add, update, or delete individual SLO criteria from the list.

Once you have finished making changes, click the "Update" button in the bottom right to save your changes.

You can also set performance change thresholds, which evaluates changes in various performance components and trigger alerts when those thresholds are exceeded.

Thresholds can be set against the following criteria:

• Change in any latency component (in milliseconds)
• Change in any latency component (as a percent of all API calls run)
• Change in pass/fail ratio (as a percent of all API calls run)

You can find these settings on the Conformance Settings page, below the SLO settings.