This includes the schema for rows in the CQ attempts table, as well as related scripts to create or update the schema.
All tables belong to a dataset, which has a name and description. Datasets can be created with the gcloud SDK bq command, which must be installed first. Datasets can be used for permission boundaries. So, to control access for different projects, the dataset used for all data directly from CQ can be put in its own dataset, whereas per-project views can be put in separate datasets.
Datasets can be created with command that look like:
bq --location=US mk --dataset --description "Attempts" commit-queue:raw
Schemas can be updated with bqschemaupdater. You should make sure you run an up-to-date bqschemaupdater:
go install go.chromium.org/luci/tools/cmd/bqschemaupdater
This tool takes a proto message and table and updates the schema based on the compiled proto message. To compile the proto message and update the schema:
go generate bqschemaupdater -message bigquery.Attempt -table commit-queue.raw.attempts