Skip to content

Observability Platform Integration

Telemetry Ingestion

Summary

Telemetry required to operate a service is shipped to the organisation's established observability platform.

Standards

  1. Telemetry required to operate a service MUST be shipped to the organisation's centralised observability platform. std-ops-telemetry-ingestion-01

Implements These Principles

Telemetry Transport

Summary

Telemetry uses an open transport, records delivery failures, and remains exportable through a periodically verified non-proprietary format.

Standards

  1. Telemetry SHOULD be shipped to the observability platform using an open, vendor-neutral protocol or collection mechanism, such as the OpenTelemetry Protocol. std-ops-telemetry-transport-01
  2. A failure to ship telemetry to the observability platform SHOULD itself be recorded, such as through an internal metric, so a gap in telemetry is not mistaken for the absence of a problem. std-ops-telemetry-transport-02
  3. Telemetry held within the observability platform SHOULD remain exportable in a usable, non-proprietary format, so historical telemetry is not lost if that platform is replaced. std-ops-telemetry-transport-03
  4. The ability to export telemetry from the observability platform in a usable, non-proprietary format SHOULD be verified periodically. std-ops-telemetry-transport-04

Implements These Principles

Platform Availability

Summary

The observability platform operates independently of monitored services, with operational telemetry buffered through brief platform outages where practical.

Standards

  1. The observability platform MUST be operated with availability independent of the services it monitors, so it remains usable to diagnose an incident affecting those services. std-ops-platform-availability-01
  2. A brief unavailability of the observability platform SHOULD NOT cause permanent loss of telemetry required for incident response. std-ops-platform-availability-02
  3. A service or intermediate collector SHOULD buffer telemetry required for incident response within defined resource bounds until the observability platform becomes reachable again. std-ops-platform-availability-03

Implements These Principles

Telemetry Retention

Summary

Each category of telemetry has a defined retention period, weighed against its value, storage cost, and any compliance need.

Standards

  1. A retention period MUST be defined for each category of telemetry, proportionate to its operational and investigative value, the cost of its continued storage, and any applicable compliance requirement. std-ops-telemetry-retention-01
  2. Business-relevant telemetry SHOULD be retained for a longer period than diagnostic telemetry, so historic trends remain available for reporting purposes. std-ops-telemetry-retention-02

Implements These Principles