Skip to main content
This tutorial takes you from a new account to first data on the Logs, Traces, and Metrics pages.

What you’ll learn

  • How to create an organization and find your collector credentials
  • How the env value routes data into a source
  • How to send data from a Node.js service or any OpenTelemetry source
  • What first data looks like and how long it takes
  • How to ask an AI assistant about the data

Prerequisites

  • A Sherlock account
  • A service to instrument, or an OpenTelemetry Collector you can configure

Step 1: Create an organization

  1. Sign in to Sherlock.
  2. If you are not in an organization yet, the app asks you to create one. To add another, open the organization switcher at the top of the sidebar and choose Create organization.
  3. Enter a name and click Create organization. Sherlock provisions the organization’s database.
The Create organization dialog with a name field

Create an organization

Step 2: Get your collector credentials

  1. Open Settings and then Collector.
  2. In the Collector Credentials card, copy the Endpoint.
  3. Click Reveal next to Bearer Token, then copy it.
Every exporter that sends to Sherlock needs both values. Keep the token in a secret store.
The Collector page, with the Endpoint field, a copy button, and the masked Bearer Token with a reveal button

Settings → Collector → Collector Credentials

Step 3: Choose the env value

Sherlock routes incoming data into a source by the env resource attribute. Decide one value per deployment of a service, for example prod or staging, and set it on every exporter:
The empty state of the Logs, Traces, and Metrics pages shows this attribute next to the endpoint and the token until data arrives.
Setup steps: configure the OTLP logs endpoint, add the bearer token, tag the env

The setup steps on an empty Logs page

Step 4: Send data

Install the Sherlock SDK, register it, and set three environment variables. The Node.js quickstart walks through it in ten minutes.

Step 5: Open the data

  • Traces shows a service within seconds of its first request.
  • Metrics lists a metric after the sender’s first export interval. The catalog refreshes every 60 seconds.
  • Logs shows a log line within about a minute. Logs need their own sender: the Node.js SDK does not ship them. See Ship and correlate logs or send them through a collector.

Step 6: Ask about it

Sherlock has an MCP server, so an AI client can answer questions over the same data. Connect Claude Code with one command and ask:
  • “Tell me about the state of the system using the Sherlock MCP tools”
  • “Group today’s errors and tell me which one is new”
The setup for the Claude apps, Cursor, VS Code, and Codex is in Ask an AI assistant.

Troubleshooting

Check the endpoint and the token against Settings → Collector. Then check the sender’s own diagnostics for a 401 or a connection error.
Compare the env value on the sender with the source’s match value. A value that matches no source is not shown.

Send data overview

Every way to send data, and what each one needs.

Node.js quickstart

First data from a Node.js service.

Go

Use the upstream OpenTelemetry Go SDK with one bootstrap file that sets what Sherlock expects. The Go code setup shows the whole file.

Ask an AI assistant

Connect an MCP client and ask questions over your data.

Alerts

Get notified when a query’s condition holds.