OpenTelemetry in, one view out
Find the slow span before the on-call does.
Lattice Telemetry lines up traces, logs and metrics by request, so one click from an alert shows which service, which query and which line.
One request, followed through every service.
Span timings, errors and tags point to the slow hop, the bad deploy or the noisy neighbour. Open the trace from a log line, a chart or an alert, and you land in the same place.
- Structured log search
- Search logs by field with a typed query, and jump from a log line straight to its trace.
- Alerts with quiet hours
- Routing rules per team, so the right person hears about the right problem and nobody else wakes up.
- OpenTelemetry from the start
- Send OTLP from any language. No agent of ours to install or keep up to date.
- Service maps
- Which service calls which, drawn from live traffic instead of last year's diagram.
From the first span to on-call
-
1
Point your collector at us
Add one exporter to the OpenTelemetry collector you already run, or follow the SDK guide for your language.
-
2
Watch traces arrive
Traces, logs and metrics line up by request as they land, and the service map fills in.
-
3
Hand it to on-call
Set alert rules and routing per team, so each alert links to the trace that explains it.
Questions engineers ask first
Do we have to replace our current agents?
No. If your services already send OpenTelemetry data, you only change where the collector exports to.
Which languages are supported?
Any language with an OpenTelemetry SDK. The docs walk through setup for the common ones.
How do we keep sensitive data out?
Set redaction rules for fields like emails or tokens. They run before data is stored.
Can we try it on our own services?
Ask for a trial when you book a demo and tell us which services you'd like to start with.
Bring one slow endpoint to the demo.
Ask to see it traced on your own services.