M3 vs other time series databases

M3 is a time series data platform designed to handle massive amounts of metrics scalability needs, as experienced by Uber. It consists of the M3 Coordinator, M3DB (a native distributed TSDB), and M3 Query Engine. The M3 Coordinator serves as a bridge between collection agents and storage in M3DB, while the M3 Aggregator handles stream-based downsampling of data before passing it to M3DB for storage. M3DB is optimized for storing time series data at scale, offering enormous storage and financial savings through compression work. M3Query performs just-in-time compression in its querying algorithm to reduce memory usage and works with a columnar data format in M3DB, allowing parallel execution of queries. M3 supports PromQL (from Prometheus) and Graphite functions for querying languages but is also working on releasing its own querying language (M3QL).