knowledge-technology

What problem does it solve?

This Skill solves the problem of mismatching source-system metric names to Tsuga metric identifiers, by providing per-technology reference bundles with exact Tsuga metric names, incident shapes, derived signals, and log patterns so you can query Tsuga reliably instead of guessing.

Core Features & Use Cases

• Per-technology reference bundles: Each supported technology has its own folder under references/ with an upstream README, a complete metrics inventory (including the exact tsuga_metric_name), and a queries guide with incident shapes, derived signals, and log patterns.
• Tsuga metric-name correctness: Enforces using tsuga_metric_name from each tech’s metrics.csv for all tsuga aggregation / logs / traces queries, including AWS metrics (e.g., aws_rds_cpu_utilization) where CloudWatch source names do not work directly.
• Investigation and query composition guidance: Provides generic investigation rules (bad window vs good control window) and concrete patterns for composing tsuga aggregation queries using the required schema and time-range/metric fields.

Quick Start

Use the knowledge-technology skill to find the exact tsuga_metric_name for a tech (such as Postgres) from its metrics.csv, then run tsuga aggregation queries using that metric and a defined incident time window.