← Back to Cookbook

Monitor your OpenCode sessions with Sentry

Add the opencode-sentry-monitor plugin to get full visibility into tool calls, token usage, and model costs across every OpenCode session.

Jump to recipe
Features
AI Observability
SDKs
Node.js
Category Monitoring
Share
Time
10 minutes
Difficulty
Beginner
Steps
5 steps

Before you start

SDKs & packages
  • OpenCode installed and configured
  • Node.js 18+ installed
Accounts & access
Knowledge
  • Basic familiarity with OpenCode configuration files

1
Create a Sentry project

In Sentry, create a new project for your OpenCode monitoring data. Go to Settings → Projects and click Create Project. Select Node.js as the platform, give it a name like opencode, and copy the DSN from the project settings — you'll need it in the next step.

Creating a Sentry project

2
Install the plugin

The opencode-sentry-monitor package is an OpenCode plugin that automatically instruments your sessions. Add it to your OpenCode configuration file (~/.config/opencode/opencode.json or the project-level opencode.json).

opencode-sentry-monitor on npm 
{
  "plugin": ["opencode-sentry-monitor"]
}

3
Configure your DSN and options

Create a config file at .opencode/sentry-monitor.json (or ~/.config/opencode/sentry-monitor.json for a global setup). Add your DSN and set tracesSampleRate to 1 to capture everything during setup. You can optionally enable recordInputs and recordOutputs to capture tool inputs and outputs in your spans.

{
  "dsn": "https://<your-dsn>@o<org>.ingest.sentry.io/<project-id>",
  "tracesSampleRate": 1,
  "recordInputs": true,
  "recordOutputs": true
}

4
Run OpenCode and generate activity

Start an OpenCode session by running opencode in your terminal and ask it to do a few things — read some files, run a command, write some code. The more tool calls it makes, the richer the data you'll see in Sentry. Every session will be captured as a gen_ai.invoke_agent span, with each tool execution — bash, read, grep, and others — tracked as a gen_ai.execute_tool child span. Token usage and model details are recorded automatically.

5
Explore tool calls and token usage in Sentry

Head to AI Agents Insights in Sentry. You'll see LLM calls broken down by model, total tokens consumed, tool call volume, and per-session traces with cost estimates. Click any trace to drill into every tool call and message in that session.

AI Monitoring documentation
Sentry AI Agents dashboard showing LLM calls by model, tokens used, tool calls, and a trace list with per-session cost and duration data from OpenCode sessions

That's it.

Every tool call, tracked.

You get a complete picture of how your AI coding agent uses models and tools — so you can tune it, debug it, and understand exactly what it's doing.

  • Installed the opencode-sentry-monitor plugin into OpenCode
  • Connected OpenCode sessions to Sentry AI Observability
  • Tracked tool calls, LLM calls, token usage, and costs per session
  • Explored per-session traces in the Sentry AI Agents dashboard

Pro tips

  • 💡 You may want to adjust tracesSampleRate based on your traffic volume and how much from your sessions you want to send to Sentry.
  • 💡 Use recordInputs: false and recordOutputs: false if your tool calls may contain secrets or sensitive file content.
  • 💡 Filter the AI Agents dashboard by agent name to isolate OpenCode traces from other AI workloads in the same Sentry project.
  • 💡 Environment variables OPENCODE_SENTRY_DSN and OPENCODE_SENTRY_TRACES_SAMPLE_RATE override the config file — useful for CI or shared machines.

Common pitfalls

  • ⚠️ Forgetting to restart OpenCode after adding the plugin — it won't pick up the new config until the session is restarted.
  • ⚠️ Using a Python or frontend Sentry DSN instead of a Node.js project — the plugin uses @sentry/node and needs a server-side project.
  • ⚠️ Setting tracesSampleRate: 0 by mistake — this silently disables all tracing, so no data appears in Sentry.
  • ⚠️ Placing the config file in the wrong location — the plugin searches .opencode/sentry-monitor.json first, then ~/.config/opencode/sentry-monitor.json.

Frequently asked questions

Yes. The plugin instruments OpenCode's session and tool layer, so it captures data regardless of which model provider (OpenAI, Anthropic, Gemini, etc.) you have configured in OpenCode.
No. Sentry uses asynchronous, non-blocking transport. Spans are batched and sent in the background with negligible performance impact.
Traces are billed based on volume. With tracesSampleRate: 1, a heavy day of coding might generate a few hundred traces. Sentry's free tier includes 10,000 spans/month — more than enough to get started.
Only if you enable recordInputs and recordOutputs as shown in the config above. With those disabled, only span metadata — tool names, durations, token counts — is captured. No file contents or prompt text is sent.
Yes. Each developer configures the plugin with the same Sentry DSN to funnel all sessions into one project, giving you a combined view of AI agent usage across the team.

What's next?

Set up alerts on token spend

Create a Sentry alert to notify you when token usage or cost spikes unexpectedly.
Go to Alerts →

Explore full trace details

Use the Trace Explorer to drill into individual OpenCode sessions and see every LLM call and tool invocation.
Open Trace Explorer →

AI Monitoring docs

Learn more about how Sentry tracks token usage, model costs, and tool calls for AI applications.
Read the AI Monitoring docs →

Fix it, don't observe it.

Get started with the only application monitoring platform that empowers developers to fix application problems without compromising on velocity.

Try Sentry for Free Explore our Sandbox