Cloud Monitoring
Start an investigation from a Cloud Monitoring alert, with the alert, its policy and its metric window as cited evidence.
Read the alert and its metric
When a Cloud Monitoring incident opens, Proofline starts an investigation. The notification is only the trigger: the investigation reads the alert itself, with its open and close times, and its policy’s conditions, such as each threshold’s comparison, value and duration. It keeps no filter text, display name or policy documentation.
It then reads the alert’s metric on the alerting resource in one-minute buckets, from 30 minutes before the alert opened to at most 30 minutes after. A read stops at one day, 16 pages, 1,024 metric rows or 1 MiB of responses, and keeps no token, provider error text or log content. Kian retains each read as evidence for that investigation, cited by its query and window.
Investigate an alert that follows a deploy
Suppose a Cloud Run 5xx alert opens soon after a release. The investigation sets the alert’s request-count series beside the environment’s deploys, merges and configuration changes. A deploy that landed after the symptom began and before the alert opened ranks above the other leads.
When a policy compares a fixed threshold, the investigation reproduces it over the read’s buckets. A bucket that crosses it shows that one value crossed, without showing how many buckets the alert needed or why the value rose. An empty read establishes neither zero activity nor a healthy service.
Route alerts to Proofline
Proofline receives Cloud Monitoring notifications through a webhook channel with Basic authentication or through a Pub/Sub subscription. Proofline sets up each intake for one organization, environment and Google Cloud project, and its reads use Monitoring Viewer access to that project, refusing any resource outside the alert’s labels. Alert intake is not self-serve, and alarm investigations are not available to new workspaces today. Talk to the team about availability.
Find an environment’s investigations
Get started with GitHub