Telemetry
About Telemetry

Structured events that agents can install and humans can review

Telemetry is a focused event analytics product for software teams. Applications and coding agents send pragmatic JSON events; teams inspect the resulting tables with SQL, charts, dashboards, exports, and alerts.

A direct event model

A named workflow becomes a table with stable fields. The same event can support debugging, operational reporting, and product analysis without hiding the underlying rows.

Reviewable analysis

Queries remain visible as SQL. Every public recipe includes the expected schema, synthetic output, interpretation, edge cases, and a dashboard or alert recommendation.

Safety by default

The guidance favors categories and identifiers over raw payloads. Prompts, credentials, request bodies, and private customer content stay out unless a team explicitly approves them.

Editorial and technical review

What “reviewed” means on this site

Product documentation is checked against the public API contract and the behavior represented in this repository. SQL recipes are restricted to read-only statements and can be planned against the supported DataFusion dialect with the optional DataFusion audit.

Example rows and charts are clearly labeled synthetic. They demonstrate an analysis pattern rather than customer benchmarks or guaranteed results. Comparison pages link to the alternative product’s own documentation and include a visible review date.

Material changes update the page-specific review date. We do not change dates merely to make unchanged content look new.

Inspect the product through a real question

Start with a SQL recipe, review the event contract, then adapt the query to a workflow in your application.