Monarch is a globally distributed, in-memory, time series database developed by Google to monitor its other billion-user applications and services. It ingests and stores data regionally; but also features global query and configuration planes to facilitate the analysis of geographically distributed data and enable unified control, respectively. To rapidly detect and mitigate outages, Monarch trades off consistency for high availability and partition tolerance, to enable the fast detection and mitigation of outages. Its data model consists of “key” columns drawn from the target (job) and metric schema, and a value column representing the time stamped value of the given metric. Its distributed architecture is localized such that data is routed according to location fields, and it includes an expressive query language equipped with field hints to speed up lookups and updates.