Skip to main content

​
Overview

Vercel log drains forward your deployment logs to Squasher over HTTPS. Squasher filters for error and fatal entries, groups them, and runs AI triage automatically. This is the fastest way to get started — no SDK, no code changes, no redeployment.

​
Setup

1

Get your drain URL

Find your project ID in the Squasher dashboard under Settings > General.Your drain URL is:
https://ingest.squasher.ai/v1/drain/vercel/{your_project_id}
2

Add the drain in Vercel

  1. Go to your Vercel project Settings > Log Drains
  2. Click Add Log Drain
  3. Select Custom as the destination
  4. Enter your drain URL
  5. Select NDJSON as the format
  6. Under Sources, select at least lambda and edge
  7. Under Environments, select production (and optionally preview)
  8. Click Create
3

Verify

Trigger an error in your Vercel deployment. Within seconds, it appears in your Squasher dashboard with an AI triage summary.

​
What Gets Captured

Squasher ingests all log entries but only creates error groups from:
  • Entries with level set to error or fatal
  • Entries with statusCode >= 500
  • Entries with statusCode = -1 (lambda crash, no response returned)
All other entries (info, warning, build logs) are stored for context but don’t trigger alerts.

​
Log Entry Fields

Squasher processes these fields from each Vercel log entry:
FieldDescription
messageLog message content
levelSeverity: info, warning, error, fatal
sourceOrigin: build, lambda, edge, static
statusCodeHTTP response status code
pathRequest path
deploymentIdVercel deployment identifier
environmentproduction or preview
requestIdUnique request ID for tracing

​
Combining with SDK

You can use both a log drain and the SDK simultaneously. The SDK provides richer error context (stack frames, breadcrumbs, user context) while the log drain catches anything the SDK misses (infrastructure errors, build failures, edge function crashes). Squasher deduplicates errors using fingerprinting, so the same error won’t be counted twice.