SchemaPulse

SchemaPulse is a web app (with a lightweight agent) that monitors time-series data shape and quality across common TSDBs. It detects “silent” breaking changes—new/vanished tags, exploding cardinality, unit changes, sampling-rate shifts, and field type flips—then shows impact: which queries, alerts, and dashboards will degrade. Instead of generic anomaly detection, it focuses on metadata and query semantics: it learns your normal tag sets per metric, expected ranges for cardinality, and typical ingestion patterns. When drift occurs, it generates a concrete remediation plan (e.g., drop/rename tags, add rollups, adjust retention, fix instrumented code) and can open a PR with suggested instrumentation changes. It also provides a weekly “data contract” report for observability and IoT teams, helping prevent runaway costs and on-call incidents caused by bad metrics hygiene. Realistically, it wins by saving money and reducing alert noise, not by being flashy.