How do I filter Vercel runtime logs for a function request?

The Citead team, updated

In short. Locate a function request with dashboard filters or a CLI request identifier query, then inspect its log details.

Definition: Vercel runtime logs record function output grouped by request. A function request can be located through dashboard filters or a request identifier query in the command-line interface (CLI).

As of 17 October 2026, Vercel documents the following workflow.

Key facts

Which dashboard filters locate the request?

Vercel's runtime log filters are available in the project's Logs sidebar. A practical sequence selects the time window and environment, then Resource: Vercel Functions and Request Path. Route matches a template such as /blog/[slug]; Request Path matches /blog/my-post.

Host and Deployment narrow the domain or deployment. Request Method selects methods such as GET; Status Code selects response codes. Free-text search covers message and requestPath; other fields use structured filters.

What appears when a log is opened?

Vercel's log details documentation describes a sidebar containing Request Id, method, path, UTC time, status, user agent, search parameters, region, firewall outcome and cache status. Function details include name, location, runtime, duration, memory usage and start type. Events and outgoing requests supply execution context; the bottom panel orders log messages chronologically.

How does a request identifier query work in the CLI?

The CLI query reference uses a subset of Kibana Query Language, a syntax for filtering fields and combining conditions. The following command combines its documented request identifier filter with a production environment and an expanded time window. <id> is a placeholder for the actual identifier.

vercel logs runtime --environment production --query 'requestId:"<id>"' --since 1h

The complete query uses single quotes; identifier values containing colons require double quotes. A request identifier search still respects the time window. The default window covers the last 30 minutes, so --since matters for older requests, within the plan's retention period.

The same reference documents vercel logs runtime --help for available filter fields and --deployment for restricting a deployment. Query results return as a bounded set; the dashboard's Live mode follows incoming logs.

FAQ

Why do I see several messages for one request?

Vercel's Function Logs documentation describes separate entries for writes to standard output during streaming. Page rendering and regeneration can also produce several logs for the same request.

Can I use the older --request-id option?

For the documented CLI version, Vercel's option migration table replaces it with --query 'requestId:"<id>"'. Removed options return a usage error with a replacement.

Sources

  1. Vercel, "Runtime Logs", 28 August 2026
  2. Vercel, "vercel logs", 10 February 2026
  3. Vercel, "Vercel Function Logs", 11 August 2026

Check how a page reads to answer engines. Open the GEO score analyzer

Test a page and see what to fix first

The analysis usually takes less than 10 seconds and needs no account.

400,000+ pages scored in the public benchmark