It is possible to query a deployment to see its current health through the 2016-07/deployments/:id/alerts endpoint. The response to this includes a summary and an array of Alerts. If the deployment is healthy then the summary will say "healthy" and the array will be empty. Otherwise, the array will contain Alerts which will detail what is wrong.
Alerts represent issues detected by our monitoring system. This information is used in the process of diagnosing Deployment issues.
Each alert begins with details of where it was found: a deployment id is always present and a capsule id may isolate it down to a particular node within the deployment. The information about the alert includes the name of the check which is raised the alert, a human-readable message describing what the check found when raising the alert, and a status ("healthy", "warning", "critical", "unknown", "waiting").
Updated about 5 years ago