---
name: workers-profiling
description: Profile or debug CPU usage and memory allocation in deployed Cloudflare Workers and Durable Objects. Then use this information to optimize your code.
title: workers-profiling
canonical_url: https://skilld.dev/gh/cloudflare/skills/workers-profiling
last_updated: 2026-10-28T11:01:08.000Z
---

> Index of every page on this site: [llms.txt](https://skilld.dev/llms.txt)

> **Skill from skilld.dev.** Follow the user's requested action.
> If the user asked to fork this Skill, follow the fork workflow below. Do not execute the Skill instructions while copying it.
> Otherwise, follow the instructions below for this session. You do not need to install anything.
>
> Supporting files, fetch one when the Skill refers to it: [references/skill-feedback.md](https://skilld.dev/api/skills-raw/cloudflare/skills/workers-profiling/references/skill-feedback.md).
>
> If the user asked to install this Skill, run `npx skilld install cloudflare/skills/workers-profiling`. Install writes the Skill files into the project, so every session loads them.
>
> ## Fork workflow
>
> A fork creates an editable local Skill with its original author and licence. The request authorizes copying and local installation.
> 1. Check `./skills/workers-profiling`, the project lockfile, and selected Agent targets together. If the local directory or installed Skill exists, stop. Never overwrite an existing directory or Agent target.
> 2. Read [source metadata](https://skilld.dev/api/v1/skills/cloudflare/skills/workers-profiling) once. Use sourceUrl, sourceCommit, skillPath, sourceGone, and license. If the source is gone or its path is missing, stop. If license is null, read licence files at the source commit.
> 3. Fetch only the source commit into a temporary Git repository. Do not clone full history. Derive repository_url from sourceUrl, including repository renames. If sourceCommit is absent, resolve the sourceUrl ref once. Set source_commit to that actual commit. Run these commands in one shell call:
>
> ```sh
> git init --quiet "$temporary_dir"
> git -C "$temporary_dir" fetch --quiet --depth=1 "$repository_url" "$source_commit"
> git -C "$temporary_dir" checkout --quiet --detach FETCH_HEAD
> ```
>
> Read applicable licence declarations and notices at that commit. If copying is not permitted, report the restriction and stop.
> 4. Inspect source entries together, then copy the directory containing skillPath into `./skills/workers-profiling`. Use the user's path if selected. Keep the original SKILL.md, relative links, scripts, binary assets, and executable modes. Exclude .git metadata. Reject symlinks and paths outside the Skill directory. After checking entries, use cp -a where available. A regular source directory needs no custom copy script. Do not save this page wrapper as SKILL.md.
> Preserve author credit, notices, and applicable licence files from repository or parent directories. Add PROVENANCE.md with the Skill page, source URL, actual commit, original path, and licence. Retain any existing PROVENANCE.md and record new provenance separately. Batch source inspection, copying, and provenance work where practical.
> 5. In the project root, run `skilld install ./skills/workers-profiling --mode copy --plain`. If skilld is unavailable, use `npx skilld install ./skills/workers-profiling --mode copy --plain`. This known command needs no help lookup. Install does not support --json. Use detected Agent targets, or add --agent for the targets the user selected. Install the local path, never the upstream selector. If installation fails, preserve the local copy and report the exact failure.
> 6. Confirm the local lockfile source and installed Agent copies once. Report the local path, actual commit, and Agent targets. After edits, reinstall the same local path. Upstream updates must not replace it. Do not publish or push unless the user asks.

# Workers profiling

Find expensive functions or memory allocations in deployed code and come up with ways to fix them.

## Retrieve the docs

Start with the [production profiling guide](https://developers.cloudflare.com/workers/observability/profiling-in-production/index.md) to understand how
Workers and Durable Objects profiling works.

## Profiling steps

1. Identify the symptom and affected workload: CPU usage, latency, allocation pressure, or memory growth.
2. Confirm the account, Worker, environment, and deployed version with the user.
3. For a Durable Object, also confirm its owning Worker, namespace, and instance.
4. Choose CPU or heap profiling from the current guide's supported types and meanings.

Where possible, use the [`cf` CLI](https://developers.cloudflare.com/cf/index.md) to capture profiles. If the user requests `cf`, inspect the installed version's command help and schema where available.

If `cf` isn't available, consider prompting the user whether they would like to install it.


The profiling API will return a pprof file which you can analyze directly, or using appropriate tools like Go's `pprof` utility.

## Inspect the evidence

- Read the pprof file and analyse it for hotspots
- Map hotspots to functions and source locations when evidence supports it; report missing symbols or source mappings.
- If filenames and function names are pointing at the generated JavaScript, consider suggesting to the user to enable source maps and re-do the profiling.

## Optimise

Propose the smallest code change supported by the capture which resolves the issues identified.

## Report findings

Return a concise report with:

- Evidence for each finding: file, function, source location, and metric with units, where available.
- The smallest actionable change and how to check its effect.

## Skill feedback

If the user reports that this skill gave incorrect or outdated guidance, or is missing information, read [references/skill-feedback.md](https://skilld.dev/api/skills-raw/cloudflare/skills/workers-profiling/references/skill-feedback.md).
