knowledge-technology

Per-technology reference bundles map source-system metrics to exact Tsuga metric names from each tech's metrics.csv, ensuring reliable incident triage instead of guessing identifiers. You avoid query failures by using the mapped Tsuga names for all aggregation, logs, and traces queries.

Tsuga aggregation query composition follows ready-to-use CLI patterns that honor the Tsuga schema, bounded time-range filters, and bad-window versus control-window investigation. Each technology bundle provides concrete incident shapes and derived signals to guide correct queries.

AWS CloudWatch source names fail because Tsuga requires its own metric identifiers, such as aws_rds_cpu_utilization, instead of native source-system names. You must use the exact tsuga_metric_name from the technology's metrics.csv to execute valid aggregation, logs, and traces queries.

Yes, per-technology reference bundles cover approximately 35 technologies including Postgres, providing complete metrics inventories, incident shapes, derived signals, and log patterns. Each technology has a dedicated folder with an upstream README and a queries guide.

Investigation uses generic rules comparing the bad incident window against a good control window to isolate anomalies. The reference bundles provide concrete query composition patterns that apply bounded time-range filters to Tsuga aggregation metrics for accurate triage.

No, native monitor or source-system names cannot be used to query Tsuga metrics directly. You must look up the exact tsuga_metric_name from the relevant technology's metrics.csv file and use that identifier for all subsequent Tsuga aggregation, logs, and traces queries.