Telemetry: Cloud Storage
In order to reduce the size of the Chromium source checkout, Telemetry stores binaries in Cloud Storage. This includes:
- Recordings of webpages, known as "page set archives" or "WPR archives."
- Support binaries, such as device/host_forwarder, ipfw, minidump_stackwalk, crash_service.
- credentials.json, which is in Cloud Storage to provide access rights.
Many benchmarks require these files to run, and will fail without them.
When recording performance tests, we generally don't call this script directly. Instead, we use update_wpr.
Set Up Cloud Storage
Install depot_tools
Follow these instructions to install depot_tools.
Request Access (for Google partners)
Many of the page set archives are in the chrome-partner-telemetry bucket,
which is not accessible publicly. Google partners can request access by emailing
telemetry@chromium.org or submitting a ticket here.
Authenticate into Cloud Storage
Some files in Cloud Storage include data internal to Google or its partners. To run benchmarks that rely on this data, you need to authenticate. Run the command below and follow the instructions for authentication with your corporate account.
$ python depot_tools/gsutil.py config
When prompted with “What is your project-id?”, just enter 0. Note that this
command is in depot_tools, not part of the Google Cloud SDK.
Buckets
Telemetry has three Cloud Storage buckets you can put binary data in.
cloud_storage.PUBLIC_BUCKET == 'chromium-telemetry'cloud_storage.PARTNER_BUCKET == 'chrome-partner-telemetry'cloud_storage.INTERNAL_BUCKET == 'chrome-telemetry'
BSD-compatible support binaries should all go into PUBLIC_BUCKET so that
everyone can use them.
Google wants to avoid legal issues with distributing third-party content, so to
be safe, most recordings of websites on the public web go in
PARTNER_BUCKET, which is accessible by Googlers and certain Google partners.
Recordings of Google-properties on the public web can go in PUBLIC_BUCKET, and
recordings of unreleased or internal Google websites go in INTERNAL_BUCKET.
Upload to Cloud Storage
Upload your files into the bucket “chromium-telemetry”
Put the target file in the directory you want it to be when downloaded from
Cloud Storage, say path/to/target. Use this command to upload:
$ python depot_tools/upload_to_google_storage.py --bucket chromium-telemetry path/to/target
A SHA1 file path/to/target.sha1 will be generated for each uploaded file.
Check the .sha1 files into the repository
$ git add path/to/target.sha1
Download from Cloud Storage
Download the file in Python
from telemetry.util import cloud_storagecloud_storage.GetIfChanged('path/to/target', cloud_storage.PUBLIC_BUCKET)
Download the file manually
$ python depot_tools/download_from_google_storage.py -s [target.sha1] -b chromium-telemetry