Developer-first API observability

Understand every request without the observability overhead.

PulseLog is a lightweight, multi-tenant request monitoring and error-tracking engine. Ingest structured events, inspect failures, and monitor latency from one focused dashboard.

Create a workspace Read the documentation

Edge ingestionShareable URL filters14-day retention

Send an event

JavaScript · Edge ingestion

HTTP 202

Built for debugging

The signal you need, without the noise.

Every screen is backed by workspace-scoped Neon queries. There is no sample traffic in the dashboard.

Real request analytics

Track request volume, p95 latency, and 5xx error rate across the latest 24 hours.

Searchable event stream

Filter by status, route, time range, service, errors, or deeply nested JSON metadata.

Full request context

Open any event to inspect its error message, stack trace, timing, and formatted JSON payload.

Isolated workspaces

Better Auth sessions and ownership-scoped queries keep every workspace and credential private.

Create a workspace

Issue an API key

Generate a pl_live_ credential. Copy it once; only its SHA-256 hash is stored.

Send events

POST request telemetry to the Edge ingestion endpoint and explore it immediately.

Documentation

Get PulseLog running

This guide covers the complete path from creating an account to ingesting, filtering, and retaining production events.

Quick start

Step 1

Create an account

Step 2

Complete onboarding

Choose a use case and create a private workspace.

Step 3

Generate a key

Open API Keys, create a named credential, and copy it once.

Step 4

Send telemetry

Call the ingest endpoint and refresh Overview or Logs.

Ingest API

Send a JSON object to POST /api/v1/ingest. Include the raw API key in the x-api-key header. The key determines the destination workspace.

Request body

Content-Type: application/json

Field	Type	Required	Description
service_name	string	Yes	Service emitting the event; 1–255 characters.
route	string	Yes	Normalized request route; 1–255 characters.
status_code	integer	Yes	HTTP status between 100 and 599.
duration_ms	integer	Yes	Non-negative request duration in milliseconds.
error_message	string \| null	No	Error summary, truncated to 2,000 characters.
metadata	JSON object	No	Structured context with an 8 KB serialized limit.

JavaScript example

Accepted response

HTTP 202

Payload guardrails

Metadata over 8,192 bytes is rejected with HTTP 413. Error messages are safely truncated to 2,000 Unicode characters before storage.

Dashboard and logs

Overview calculates request count, p95 latency, 5xx error rate, and hourly 2xx/4xx/5xx buckets from the latest 24 hours. Logs are ordered newest first and limited to 100 rows per query.

Shareable log filters

Filters are encoded in the URL and evaluated by the server.

Parameter	Example	Behavior
status	500 or 5xx	Exact status code or the 2xx, 4xx, and 5xx classes.
route	/api/orders	Exact route match.
q	timeout	Search service, route, error message, and JSON metadata.
range	15m	One of 15m, 1h, 24h, or 7d. Defaults to 24h.

/<workspaceId>/logs?status=5xx&route=%2Fapi%2Forders&q=timeout&range=24h

Security and retention

Tenant isolation

Every dashboard read and mutation validates the Better Auth user against the workspace owner. API keys can ingest only into their associated workspace.

One-time credentials

API keys contain 192 bits of random entropy. Only a SHA-256 hash is stored, and revocation deletes the credential immediately.

Automatic retention

The cleanup job removes events older than 14 days and retains no more than the newest 50,000 events in each workspace.

Scheduled cleanup

Vercel invokes the protected cleanup endpoint daily at 03:00 UTC using the CRON_SECRET bearer token.

Ready to inspect your first request?

Create a workspace, issue a key, and start sending structured events.

Get started