Capturing Metrics

Azure AI SDK automatically captures relevant invocation and usage metrics from each API call and logs them to Statsig. You can see these events streaming in real-time in the console at: https://console.statsig.com/metrics/events.

Metrics captured

Completion invocations capture the following metrics automatically: completion token length, prompt token length, latency, model name, total token length.

For example, a completion API call results in a usage log that looks like this:

These could be used to compare multiple deployments with each other, and also to run experiments against different sets of parameters, aiding in optimization of cost, responsiveness, user experience, etc.