What problem does it solve?
Developers writing DOCA applications often confuse the publisher-side Telemetry Exporter library with the receiving Telemetry Service, misuse the schema-to-source lifecycle, and hit DOCA_ERROR_* failures without a clear diagnosis path. This Skill gives an AI agent the exact lifecycle, publish-surface selection rules, error taxonomy, and debug workflow for the doca_telemetry_exporter_* API.
Core Features & Use Cases
- Publish-surface selection: Chooses between typed structured events, opaque events, the Metrics API (counter/gauge/histogram), OTLP logs, and the NetFlow/IPFIX sibling API based on user intent.
- Lifecycle and error guidance: Walks the schema_init → schema_start → source_create → source_start → report → flush → destroy sequence and maps DOCA_ERROR_BAD_STATE, INVALID_VALUE, NO_MEMORY, INITIALIZATION, and UNKNOWN to root causes.
- Task workflows: Provides configure, build, modify-from-shipped-sample, run, test, and debug verbs grounded in the installed DOCA samples under /opt/mellanox/doca/samples/doca_telemetry_exporter/.
- Use Case: A developer wants their DOCA Flow program to publish per-second packet counters to a downstream collector; the agent identifies the Metrics API as the right surface, starts from the telemetry_export_metrics sample, and validates delivery with a file-write smoke test.
Quick Start
Ask your agent to help you emit a structured event or counter from your DOCA application using the DOCA Telemetry Exporter library on your installed DOCA host.