Back to websiteSample workspace with illustrative data. Start a free trial to see your own.Sign in
Sample data

Services

Connect a service

All services

Showing 8 of 8 services
How these numbers are counted. A request is an entry span (server, consumer, or a root span) with health probes excluded. A failed request is a distinct trace with an error span in that service, never more than the requests. Window: the last 60 minutes, ending 14:07 GMT. Overview, Query Studio, Alert Rules and AIOps use the same rule.
Updated , just now Auto-refresh every 30 s
Status
Matches name, role or status. The list updates as you type.
Services in the last 60 minutes ending 14:07 GMT. Press Enter on a row to open its details.
Service Status Requests Failed p95 Failed trend Last seen Open details

How status is decided. Failing: 5% or more of requests failed. Degraded: 1% to under 5% failed, or p95 over a latency rule. Healthy: under 1% failed. Too little traffic: fewer than 10 requests, so no rate or p95 is judged. No data: no spans in the window. Failed trend is failed % per 5 minutes on one shared 0 to 10% scale, so rows compare by height. Too little traffic rows show raw failed counts.

All services

Loading services

No services are reporting yet

A service appears here as soon as Onyx receives its first span or log line. This page checks every 10 s, so you do not need to reload it.

  1. Choose how to send data in onboarding: OpenTelemetry, a log shipper, or an agent.
  2. Point your exporter at your ingest endpoint with your ingest key and a service name:
OTEL_SERVICE_NAME=your-service-name
OTEL_EXPORTER_OTLP_ENDPOINT=<your ingest endpoint>
OTEL_EXPORTER_OTLP_HEADERS=authorization=Bearer <your ingest key>
  1. Send one request through the service. It shows up here within about a minute.

A service needs at least 10 requests in 60 minutes before Onyx rates it. Until then it shows as Too little traffic, never as healthy.

Send your first event Waiting for the first event. Checking every 10 s.