mcpbeat

Download Profile Trace

microsoft/download-profile-trace

Download a profiler trace file from Application Insights Profiler. Use this when asked to download, save, or export a profiler trace or .etl/.nettrace file.

6k tokens
context cost
the whole folder, loaded on every use
6
files
ships runnable scripts
0
copies elsewhere
how many repositories repackaged it
3
stars on the repo
on the repository, not the skill itself

Install

one command, takes just this skill from the repository
npx skills add https://github.com/microsoft/code-optimizations-skills --skill download-profile-trace

What comes with it

19 333 bytes besides the instruction
references/trace-file-info.md
scripts/download-trace-by-location.md
scripts/download-trace.md
scripts/list-profile-traces.md
scripts/resolve-trace-identifiers.md

The instruction itself

11 sections, as written by the author

Download Profile Trace

This skill downloads a raw profiler trace artifact from the Application Insights Profiler dataplane API. The downloaded file can be opened in PerfView, Visual Studio, or other trace analysis tools.

Prerequisites

  • The user must be logged in to Azure CLI (az login)
  • The user must provide an Application Insights app ID (GUID) or resource ID

Steps

1–2. Check investigation notes and gather inputs

Follow the Standard Skill Preamble to check for existing investigation context and gather inputs.

In addition to the standard App ID, this skill requires:

  • Time range (optional): How far back to look for traces. Defaults to 7 days (P7D).
  • Output path (optional): Where to save the downloaded file. Defaults to the current directory.

3. Acquire an access token

Run the script in get-access-token.md to acquire a Bearer token for the profiler dataplane. See the token freshness and session scoping guidance in that document — re-acquire the token in the same command block as each API call.

4. List available profile traces

Run the script in list-profile-traces.md to call the ingested artifacts endpoint and list available profiler traces.

Display only the 10 most recent traces by default. If more traces exist, tell the user the total count and offer to show more. The API can return hundreds of traces for active applications, which produces overwhelming output.

Present the results to the user with:

  • Timestamp (triggerTime — when the profiling session was triggered)
  • Role name
  • Role instance (machine/container name)
  • Artifact ID (note if null)
  • Format (Etl, Nettrace, or Netperf)

When constructing the output filename, derive the file extension from two fields:

  • format → base extension: Etl.etl, Nettrace.nettrace, Netperf.netperf
  • blobUri → compression: if the URI ends in .zip, append .zip to the base extension

> ⚠️ Do not use the blobUri extension as-is for the filename. The blobUri may show .etl.zip even when format is Nettrace, which would produce a misleading filename. Always use format for the base extension and blobUri only to detect .zip compression.

Include the role name in the filename to avoid collisions when downloading traces from different resources into the same directory, e.g., trace-myapp-2026-03-20T214135.nettrace.zip.

Let the user pick which trace to download. If there's only one result, confirm it with the user before proceeding.

5. Download the selected trace

Choose the download method based on whether the selected trace has an artifactId:

5a. If artifactId is available (not null)

Run the script in download-trace.md with the selected artifactId. The script:

  • Calls the artifact download endpoint (GET /artifacts/{artifactId})
  • Handles the 302 redirect to get the blob download URL
  • Downloads the trace file to the specified output path
5b. If artifactId is null

First, run the script in resolve-trace-identifiers.md to query Application Insights customEvents for the ServiceProfilerSample event matching the selected trace. This can resolve:

  • Artifact ID (from ServiceProfilerContext) — if found, use the standard download in step 5a
  • Trace location ID (from ServiceProfilerContent) — if found, use the trace location download below

If an artifact ID is resolved, go back to step 5a and download using that artifact ID.

If only a trace location ID is resolved, use download-trace-by-location.md to download. The trace location ID is a pipe-delimited v1 string:

v1|{stampId}|{appId}|{machineName}|{processId}|{etlFileSessionId}

The script:

  • Calls POST /artifacts/byArtifactLocation?t={traceLocationId} to get a SAS-protected download URL
  • Downloads the trace file from the returned URL

If neither identifier is resolved from the query, ask the user for the trace location ID or artifact ID manually.

6. Present the result

After downloading, tell the user:

  • Where the file was saved
  • The file size
  • If the file is a .zip, remind the user to extract it before analysis
  • How to open it — see trace-file-info.md for guidance on tools and formats

Response format

See trace-file-info.md for information about trace file formats and how to open them.

How to use it

Copy the folder

Take microsoft/download-profile-trace from the repository into ~/.claude/skills for personal use, or into .claude/skills inside a project.

Check the name does not clash

The agent identifies a skill by the name field in its header. Two skills with the same name cannot sit side by side — one of them will be ignored.