Service overview

The Services page lists every service that has reported traces, with the RED metrics and latency percentiles for each over the chosen time range.

Title: Services · Service health and dependency overview. Refresh button top-right.

Time picker

Presets: 15m | 1h | 6h | 24h | 7d. Plus explicit From/To datetime inputs.

Source scope

The page respects the source switcher in the left rail. Pick a source to scope to traces from that source only, or use All sources to see every service.

Results table

Columns:

ColumnMeans
ServiceService name (the OTLP service.name).
RequestsTotal span count over the time range.
ErrorsCount of spans whose status code is Error.
Error RateErrors as a percent of total requests. Green when zero, amber/red when non-zero.
AvgMean span duration.
P50Median span duration.
P9090th-percentile duration.
P9999th-percentile duration.
(last column)Metrics link — opens the Metrics page filtered to this service.

When no services have reported, the table area shows:

> "No services found. Ingest trace data to see service metrics."

Service detail

Click any service row to open /services/<name>:

  • Six summary tiles — Requests, Errors, Error rate, Avg, p90, p99 — with severity coloring on the error tiles.
  • Two time-series charts — Requests + Errors over time, and p50 / p90 / p99 latency in milliseconds. Time bucket size is 60 seconds.
  • Top operations table — root span name, request count, error count, error rate, avg, p99. Up to 10 rows.
  • Recent traces table — errors first, then most recent. Click any trace ID to open the trace detail. Up to 20 rows.

The detail page respects the global time picker and the source switcher in the left rail.

Sending traces

Same OpenTelemetry setup as Traces:

  • Set OTEL_SERVICE_NAME per service (this becomes the row name).
  • Point the OTLP exporter at the agent's receiver (4317 for gRPC, 4318 for HTTP).
  • Each span counts toward Requests; spans with status Error count toward Errors.
  • Traces — the underlying span data.
  • Service Map — visualisation of service-to-service calls.