octri_monitoring (Dart)
Error and performance monitoring for Dart applications. Report a thrown object with its stack trace, emit your own application events, time spans into a request waterfall, and continue a W3C distributed trace that started in whichever client called you.
Octri turns an OpenAPI spec into a documentation site, client SDKs for ten languages, an MCP server your AI assistant can call, and monitoring for the API behind them. This package is the Dart monitoring runtime, and it works on its own: a generated Octri API SDK is not required. See octri.dev/monitoring.
Dart 3.2 or newer, on the Dart VM. No runtime dependencies.
Install
dart pub add octri_monitoring
Setup
Call init once at startup.
import 'dart:io';
import 'package:octri_monitoring/octri_monitoring.dart';
Octri.init(OctriConfig(
url: 'https://monitoring.example.com', // your monitoring base URL
token: Platform.environment['OCTRI_TOKEN'], // your project ingest token
environment: '<your project id>', // the dashboard project id
release: Platform.environment['GIT_SHA'], // optional
));
Hosted users copy the project-scoped URL, token, and environment from the
Monitoring connection settings in the dashboard. Pass token: null only when you
point at an open self-hosted ingest endpoint.
Every call is asynchronous, best effort, and carries an idempotency key. Transport failures are suppressed, so a monitoring outage cannot affect the process it is watching.
Application events
Octri.captureEvent(
'checkout.completed',
options: const OctriEventOptions(
level: 'info',
tags: {'region': 'eu-west', 'plan': 'growth'},
context: {'orderId': 'ord_912'},
),
);
OctriEventOptions also carries user, breadcrumbs, fingerprint,
operationId, method, path, statusCode, latencyMs, attempt, and
requestId. Supplying eventId makes a retried delivery idempotent.
Errors
try {
await handle(request);
} catch (error, stackTrace) {
Octri.captureError(
error,
stackTrace: stackTrace,
options: const OctriErrorOptions(
method: 'POST',
path: '/orders',
statusCode: 500,
),
);
}
Omitting stackTrace still reports the error, with the current stack.
Joining the caller's trace
Your generated client SDK sends traceparent: 00-<traceId>-<spanId>-01 on every
request. Read it on the way in and pass the result as trace, and the dashboard
groups the client call and the server error under one traceId: the request that
failed, beside the frame that threw.
final trace = Octri.traceFromHeader(request.headers.value('traceparent'));
Octri.captureError(error, options: OctriErrorOptions(trace: trace));
traceFromHeader(null) starts a fresh trace, so the same code path works for
traffic that arrives without a header.
Spans
A span is a completed unit of work. Report one per request to get the waterfall, and one per sub-operation to see where the time went inside it.
final started = DateTime.now();
final rows = await db.query(sql);
Octri.captureSpan(OctriSpan(
traceId: trace.traceId,
spanId: spanId,
parentSpanId: trace.parentSpanId,
name: 'orders.list',
operationId: 'listOrders',
startTime: started,
endTime: DateTime.now(),
));
Spans sharing a traceId nest by parentSpanId in the dashboard waterfall.
What gets redacted
Payloads are scrubbed on the way out, so a credential that ended up in a log line or a context object never reaches the dashboard.
Any key whose name looks like a credential (password, secret, token,
apiKey, authorization, cookie, ssn and the rest of the usual list) has
its value replaced with [redacted], at any depth. Matching ignores case and
separators, so api_key, apiKey and X-API-KEY are all the same key.
Free text is swept too: the message, an error message and its stack, and
anything else you send as a string. Bearer tokens, JWTs, card numbers and email
addresses come out as [redacted]. A card number has to pass the Luhn check
first, so an order number or a timestamp survives.
user is the exception. It is the field you fill with an identity on purpose,
so user.email is reported exactly as you set it. Credential-shaped keys inside
it are still redacted.
Add your own key names:
Octri.addScrubFields(['accountNumber', 'otp']);
Or take the payload yourself, and return null to drop the event:
Octri.setBeforeSend((payload) => payload['path'] == '/health' ? null : payload);
Redaction runs after your hook, so a hook cannot leak a credential by accident.
The rest of Octri
| Product | What it does |
|---|---|
| API Studio | Your OpenAPI spec becomes a hosted documentation site with a live request playground, editable page by page. |
| SDK Studio | The same spec becomes client libraries for ten languages, versioned and released together. |
| MCP | Your endpoints and docs become tools an AI assistant can call, generated from the same spec. |
| Monitoring | Errors, traces, uptime and releases for the API, joined to the SDK calls that reached it. |
Monitoring runtimes
More
MIT licensed.
Libraries
- octri_monitoring
- Standalone error and performance monitoring for Dart applications.