telemetry

Telemetry is the automated collection and transmission of measurements or event data from remote, inaccessible, or distributed systems to a receiving system, where the data is used for monitoring, analysis, and control.

In software engineering, telemetry refers to the operational data that applications emit to observability backends. It can include traces, metrics, logs, and related signals. This data is gathered by sensors or code-level instrumentation, carried via SDKs and exporters. Then, it’s analyzed to debug behavior, monitor performance and reliability, track usage and cost, and inform capacity planning.

In AI coding tools, telemetry has a second, narrower sense: the usage and diagnostic data the tool reports about itself. Editors and assistants like VS Code, GitHub Copilot, and Claude Code emit crash reports, error telemetry, and usage data such as session counts, token and cost totals, and tool invocations. Each tool gates that reporting behind an explicit control, such as VS Code’s opt-out telemetry.telemetryLevel setting or Claude Code’s opt-in CLAUDE_CODE_ENABLE_TELEMETRY variable, which exports metrics, events, and optional traces over OpenTelemetry to a self-hosted backend. These same signals feed LLM observability dashboards.