One pipeline for all observability signals

Vercel is renaming Log Drains to simply Drains, reflecting a broader expansion beyond log data. The updated feature now supports exporting OpenTelemetry traces, Web Analytics events, and Speed Insights metrics alongside traditional logs. This gives teams a unified pipeline for streaming observability data from Vercel into external systems they already operate.

The rationale is straightforward: logs alone rarely explain modern application behavior. Traces reveal how requests move through serverless functions, Web Analytics captures user interactions, and Speed Insights provides real-user performance data. Drains consolidate these signals so teams can send complete, correlated telemetry to tools like Datadog, Honeycomb, Grafana, Elastic, or their own data warehouses.

Drains support four data types:

  • Logs: Runtime, build, static asset, firewall, and function logs.
  • Traces: Distributed tracing data forwarded from deployments.
  • Web Analytics: Page view and custom event data.
  • Speed Insights: Real-user performance metrics and Web Vitals.

Because Vercel manages the entire request path from browser to function execution, these signals share consistent context. Logs from traced requests will automatically include traceId and spanId, allowing engineers to jump from a log line directly into its corresponding distributed trace.

Custom endpoints vs. managed integrations

Drains come in two connection forms. Custom drains let you stream data to any HTTP endpoint you control. You can choose the data type (logs, traces, analytics, or performance metrics), configure sampling rates to manage traffic volume, select JSON, NDJSON, or Protobuf serialization, and add headers or signature verification for security. This is useful for routing traces to a self-hosted OTLP collector, logs to a private Elastic cluster, or analytics events into Snowflake.

Integration drains, conversely, offer turnkey setup with vendors such as Dash0, Statsig, Datadog, and Logflare. These connections automate configuration and begin streaming data into your vendor account without manual endpoint management.

Correlated context across signals

The key advantage of routing Drains through Vercel is the inherent correlation. A spike in LCP can be traced back to a slow API call; a 500 error log can be linked to the specific request that caused it. Exporting all four signal types together provides a continuous production view within your existing tooling.

Traces follow the OpenTelemetry protocol, so they plug directly into Datadog APM, Honeycomb, Grafana Tempo, New Relic, or any OTel-compatible backend, without custom instrumentation. Log and trace correlation happens automatically, and analytics data can land in the same warehouse used for business reporting.

Availability and pricing

Drains are available now for Pro and Enterprise plans. Data exported via Drains is billed at $0.50 per GB, matching the existing log drain rate. Teams can configure multiple drains across as many projects as needed.

Setup

Configure Drains from Vercel dashboard → Team Settings → Drains. Choose a custom HTTP endpoint for any data type, or install a managed integration from the Marketplace. Configuration details, supported formats, and schema references are available in the official documentation.