Glossary · 6 · SDK interoperation
Telemetry
Also known as: Telemetry data, Remote telemetry
German: Telemetrie
In engineering and software operations, telemetry is the automated collection of measurements and event data at remote or distributed sources, such as devices, services and applications, and their transmission to another system for monitoring and analysis; in software it typically comprises metrics, logs and traces.
- SDK interoperation
- Hardware SDKs
In one sentence
Telemetry is the automated collection and transmission of measurements and events from devices and services, such as metrics, logs and traces.
Example
A fleet of drives reports temperature and load every minute to the vendor's cloud service, and the service's own API emits traces for each request it handles.
How it applies
- Hardware SDKs: Device SDKs collect sensor values, status and diagnostic events and send them to a gateway or cloud service, often over protocols such as MQTT. In the Internet of Things (IoT), device telemetry is the main data stream from the field.
- Developer SDKs: Client libraries and services emit telemetry about their own behavior: request counts, latency, errors and trace spans. Logging and distributed tracing are two forms of it.
- Interoperation: Telemetry crosses the boundaries between the SDK types: data measured on a device passes through a hardware SDK, a cloud API and a platform extension before an operator sees it. Shared formats and conventions, such as those of OpenTelemetry, let one system follow it end to end.
- SDK usage telemetry: Some SDKs and tools also report anonymous usage data to their vendor. Documentation should state what is collected, why, where it is sent and how to switch it off; such data can include personal or confidential information.
- Documentation: Telemetry fields, units, sampling rates and retention belong in the reference documentation, because operators and analysts rely on them.
Telemetry vs. observability
Telemetry is the data. Observability is the property of a system that lets people understand its internal state from that data, and the practice of achieving it. A system can produce a great deal of telemetry and still be hard to observe if the data lacks context, correlation or the right signals.
Telemetry vs. telecontrol
Telemetry carries measurements from a remote system to a central one. Telecontrol carries commands in the other direction. In industrial systems both often share the same equipment, for example a remote terminal unit (RTU), but they have different safety and security implications.