depot_tools_metrics/

README

depot_tools monitoring app

This is a simple GAE app to collect the metrics uploaded by depot_tools and store them in BigQuery. This app is only reachable when the X-AppEngine-Trusted-IP-Request header, which is set only on requests received from corp machines, is present on the request. This is a way to ensure we don't collect data from non-Googlers.

API

This app exposes two endpoints:

• /should-upload

  Returns 200 if the request comes from a corp machine, and 403 otherwise.

• /upload

  Accepts a JSON file in the format described by monitoring_logs_schema.json and writes the data to the depot_tools table in the metrics dataset of the cit-cli-metrics project.

  Returns:

  • 403 if the request comes from a non-corp machine.
  • 400 if the reported metrics are invalid.
  • 500 if there was an internal error.
  • 200 if the request succeeded.

Deployment

Updating the Schema

To update the metrics table and schema, run:

• go generate schema/gen.go
• bqschemaupdater -table cit-cli-metrics.metrics.depot_tools -message-dir schema -message schema.Metrics

If these steps modified any files in the repo, add them to a change for review. By the end of this process, running these steps on a clean checkout of the main branch should not result in further updates to the schema.

Uploading a New Version

Make sure you're on the main branch and that your working tree is clean. Run gae.py upload. If everything worked, you should see a new version with name {ID}-{commit hash} in the cit-cli-metrics dashboard.

If gae.py upload doesn't work, try gae.py upload --app-id cit-cli-metrics --app-dir .

Migrating Traffic

After uploading, navigate to the cit-cli-metrics dashboard and migrate traffic to your new version.