ABOUT

Enterprises deploying AI applications at scale face a multi-layered observability challenge: they need to monitor not just infrastructure and application health, but also AI-specific metrics like model inference latency, token throughput, GPU utilization, prompt quality, and response consistency — all while maintaining compliance and security requirements. Dynatrace solves this by extending its existing platform with AI-aware instrumentation that automatically discovers AI pipeline components, traces requests through model inference services, correlates ML infrastructure metrics with application performance, and uses its Davis AI engine to automatically identify the root cause of AI service degradation.

INTEGRATION GUIDE

1. Automatically discover and map AI service dependencies — from API gateways to embedding services to LLM inference endpoints — without manual configuration 2. Trace LLM requests end-to-end across your AI stack, correlating prompt input, token usage, model response, and downstream processing in a single distributed trace 3. Monitor GPU cluster health, inference server throughput, and memory utilization alongside application KPIs in unified dashboards 4. Detect model drift and performance regression using Davis AI's automated baseline analysis on latency, error rate, and token consumption patterns 5. Receive AI-powered root-cause analysis that correlates infrastructure events, code changes, and LLM telemetry when response quality degrades

DYNATRACE AI OBSERVABILITY

ABOUT

INTEGRATION GUIDE

TAGS