What problem does it solve? Diagnosing service performance issues requires navigating complex monitoring systems and knowing exact metric names, regions, and tag schemas. This Skill provides a structured interface to query Metrics (BytePlot / BytedTrace) monitoring data by PSM or metric name, so you can inspect QPS, latency, error rates, and dependency traffic without manually building queries. ## Core Features & Use Cases - Preset Service Queries: Run overview, downstream, or upstream query types against a PSM to get labeled sections summarizing QPS, latency, and error rate for a service and its dependencies. - Custom Metric Queries: Query any exact metric name directly with optional filters, grouping, aggregation, downsampling, and top-k selection. - Tag Discovery: Use the suggest_tagk action to discover available filterable and groupable tag keys before building filtered queries. - Use Case: When a service's error rate spikes, ask for a downstream dependency check on the PSM in a specific region to identify which downstream call is failing, then drill into the failing metric grouped by error type. ## Quick Start Ask the agent to show the QPS, latency, and error rate overview for your service PSM in a specific region such as Singapore-Central.
