See where a request failed or slowed down
Vigilon follows each request through your apps, its database queries and its outgoing calls, and keeps every trace with an error and every slow one. Built on OpenTelemetry, for Node.js and Python, with no sample rate to tune.
The whole request, step by step.
A waterfall for every kept request
Each step is a bar by start time and duration, nested under the step that called it, with the exception on the span that failed.
Database and outgoing calls included
PostgreSQL, MySQL, MongoDB and Redis queries, and outgoing HTTP calls, are traced automatically in Node.js and Python.
Across your own apps
When one of your apps calls another, the trace carries on into it, so a request that crosses two services is one trace.
Errors and slow requests always kept
Vigilon decides after the trace arrives: every trace with an error, every slow trace and every job run are kept, plus a sample of fast, successful requests.
Counted per request, not per span
Plans count requests, so a request with thirty database calls still counts once, and there is no sample rate to set in your code.
Errors linked to their trace
An error group opens on the trace of the request or job run that produced it, so you go from the alert to the failing step.
Traced from the first request.
- 01
Install the SDK
One package for Node.js or Python, or point an existing OpenTelemetry setup at Vigilon.
- 02
Start it with your app
Vigilon starts before your code and traces your framework, database clients and HTTP calls with no changes to them.
- 03
Vigilon keeps what matters
Every trace with an error or a slow response is kept, along with a sample of the fast ones.
- 04
Open any request
Find a trace from an error, an endpoint or the Traces page, and follow it step by step.
One package, or the OpenTelemetry you already run.
Install the SDK and start it with your app, or send OTLP traces from an existing OpenTelemetry setup, alone or next to the backend you use today.
Docs What Node.js tracesWhat Python tracesExisting OpenTelemetryNext.js
npm install @vigilon/node
export VIGILON_API_KEY="[YOUR_API_KEY]"
node --import @vigilon/node/register server.jsExpress, Fastify, NestJS, Next.js and plain HTTP servers are traced, with PostgreSQL, MySQL, MongoDB, Redis and outgoing HTTP calls.
Tracing in Vigilon and Sentry.
Sentry traces requests once tracing is switched on, and bills by span. This is how the two compare, with a source for every Sentry fact.
| Capability | Sentry | Vigilon |
|---|---|---|
| Request traces with database and outgoing-call spans | YESOnce tracing is on: inbound and outbound HTTP, pg, MySQL, MongoDB, ioredis [16] | YES |
| Which traces are kept | YOUR SAMPLE RATESet in the SDK; the guide says to lower it in production [9] | AFTER THE FACTEvery background job trace. For HTTP requests Vigilon handles filtering for you. |
| Spans included each month | 5MOn every plan, about 12 requests a minute at 10 spans each; then billed per span, and dropped past your budget [1] [2] | UNLIMITEDVigilon counts requests, however many spans each one has |
| What the default install turns on | ERRORSTracing, profiling and logs are switched on in code [9] | ERRORS, TRACES, ENDPOINTSOne package and one start flag |
| OpenTelemetry | BETAOTLP ingest for traces and logs is in open beta [5] | NATIVEOTLP is the ingest; the SDKs are thin OpenTelemetry layers |
- [1] Sentry pricing: spans, logs and cron monitors included on each plan: sentry.io/pricing. Checked 2026-10-25.
- [2] Sentry pricing docs: span rates, and data dropped past the budget: docs.sentry.io/pricing. Checked 2026-10-25.
- [5] Sentry OTLP ingest: docs.sentry.io/concepts/otlp. Checked 2026-10-25.
- [9] Sentry Express setup guide: docs.sentry.io/platforms/javascript/guides/express. Checked 2026-10-25.
- [16] Sentry Node.js automatic instrumentation: docs.sentry.io/platforms/javascript/guides/node/tracing/instrumentation/automatic-instrumentation. Checked 2026-10-25.
Tracing questions.
What is distributed tracing?
Distributed tracing follows one request through every step it takes: your route handler, its database queries, its outgoing calls and the other apps it reaches. Vigilon records each step as a span and shows the request as a waterfall, so you can see where it failed or what made it slow.
Does Vigilon keep every trace?
It keeps every trace with an error, every slow trace and every job run, plus a sample of fast, successful requests. It decides after the trace arrives, so there is no sample rate to set in your code, and endpoint numbers are counted from every request before any trace is dropped.
Can I send traces from my existing OpenTelemetry setup?
Yes. Point an OTLP/HTTP exporter at Vigilon with your API key, alone or as a second exporter beside your current backend. Vigilon accepts traces; metrics and logs are not stored.
Does it trace requests across services?
Yes, between your backend apps: when one app calls another with trace context, both sides land in the same trace and the request is counted once. Browser SDKs send errors only, so traces start at your backend.
Is tracing billed per span?
No. Plans count requests and job runs, not spans, so a request with many database calls counts once.
Which frameworks are traced automatically?
In Node.js: Express, Fastify, NestJS, Next.js on the Node.js runtime and plain HTTP servers, with pg, mysql, mysql2, MongoDB, Redis and outgoing HTTP. In Python: FastAPI, Flask and Django, with psycopg, MySQL drivers, Redis, PyMongo, SQLAlchemy, requests, httpx, aiohttp and boto3.
Everything else Vigilon watches.
Error tracking
Repeated errors grouped into one, with an email when a new error appears or a fixed one comes back.
Cron job monitoring
Every run of a scheduled job or queue worker is a trace, and a job that misses its schedule shows as overdue.
Endpoint monitoring
Traffic, p50 and p95 latency, error rate and status codes for every route, counted from all requests before any trace is dropped.
Anomaly detection
Error spikes, slower endpoints, error surges, missed job runs and failing job streaks, each checked against its own history. On Team and Pro.
MCP server
Claude Code, Codex, VS Code, Claude and ChatGPT read your errors, traces and job runs over a hosted, read-only MCP server.
Start free for two months.
Pick a plan, add the SDK, and Vigilon starts watching your app. Starter is free for your first two months, with no card needed.