What problem does it solve? Investigating slow requests, failed spans, and service dependencies in distributed systems requires deep knowledge of trace data structures and query syntax. This Skill provides ready-to-use DQL query patterns for analyzing spans, traces, and request flows in Dynatrace Grail, eliminating guesswork when diagnosing performance bottlenecks and failures. ## Core Features & Use Cases - Performance Analysis: Find slow requests, compute p95/p99 latency percentiles by endpoint, and build response-time timeseries from span data. - Failure Investigation: Break down failures by reason (HTTP codes, gRPC status, exceptions), analyze exception stack traces, and list recent failed requests with trace exemplars. - Service Dependency Mapping: Trace service-to-service communication, outgoing HTTP calls, database queries, messaging flows (Kafka, RabbitMQ, SQS), RPC/gRPC calls, and serverless function invocations. - Use Case: When p99 latency spikes on a checkout endpoint, use the slow-trace detection pattern to find requests over 5 seconds, then drill into database spans with extrapolated counts to identify the slow query causing the bottleneck. ## Quick Start Ask the AI to find all traces with requests slower than 5 seconds in the last 2 hours and show the failing endpoints with example trace IDs.
