M3
What it is
M3 is an open-source, Prometheus-compatible metrics engine for scaling Prometheus monitoring and centralizing metric storage. Its M3DB component is a distributed time-series database for real-time metrics with long retention. M3 advertises 100% PromQL compatibility and supports Graphite query languages, while also listing M3QL. Ingestion options include Prometheus remote write, StatsD, Carbon plaintext, InfluxDB line protocol, and JSON. Documentation describes integrations with Prometheus, Graphite, Grafana, and InfluxDB. M3 can be installed from binaries or Docker, with Kubernetes cluster deployment covered in its documentation. The project supports horizontal scaling and says it has stored tens of billions of active metric time series. Its design uses three data replicas with quorum writes and reads, and the project reports an 11X compression ratio. M3 is released under Apache 2, with contributions welcomed. The open-source plan costs nothing. M3DB is not intended as a general-purpose time-series database, and large-scale backfilling is not currently possible.
For and against
For
- PromQL compatibility advertised at 100%
- Supports Prometheus, StatsD, and Carbon ingestion
- Integrates with Grafana and InfluxDB
- Kubernetes cluster deployment documented
- Apache 2 open-source project
Against
- Not intended as a general-purpose time-series database
- Large-scale backfilling is not currently possible
- Stored values are optimized for float64
- No commercial plan is stated
The verdict
M3 offers an open-source path to scalable Prometheus-compatible storage with multiple ingestion and query options. Review its backfilling and data-type limits before using M3DB for broader time-series workloads.
Who it suits
M3 suits teams operating Prometheus-based monitoring that need centralized, scalable metrics storage and long retention. It fits Linux, macOS, API, and self-hosted deployments.
Plans and prices
M3 (open source)
Open-source metrics engine; no commercial plan or usage limits stated
Read 11 Oct 2026 · m3db.io
On the record
- Compression
- The home page reports an optimized compression algorithm with an 11X compression ratio.m3db.io · read 11 Oct 2026
- Deployment
- M3 can be installed with binaries or Docker, and its documentation includes cluster deployment with Kubernetes.m3db.io · read 11 Oct 2026
- Ingestion formats
- M3 supports Prometheus, StatsD, and Carbon ingestion formats.m3db.io · read 11 Oct 2026
- Integrations
- The documentation describes integrations with Prometheus, Graphite, Grafana, and InfluxDB.m3db.io · read 11 Oct 2026
- Limitations
- M3DB is not intended as a general-purpose time-series database; large-scale backfilling is not currently possible and stored values are optimized for float64.m3db.io · read 11 Oct 2026
- Open source
- M3 is open-sourced under the Apache 2 license, and contributions are welcomed.m3db.io · read 11 Oct 2026
- Origin
- M3 began as an open-source metrics platform developed by Uber’s Observability team in 2014.m3db.io · read 11 Oct 2026
- Purpose
- M3 is a Prometheus-compatible metrics engine for scaling Prometheus-based monitoring and centralized metrics storage.m3db.io · read 11 Oct 2026
- Query compatibility
- M3 advertises 100% PromQL compatibility and supports Graphite query languages.m3db.io · read 11 Oct 2026
- Reliability
- The home page describes three data replicas with quorum writes and reads for consistency.m3db.io · read 11 Oct 2026
- Scale
- The home page says M3 has been used to store tens of billions of active metric time series.m3db.io · read 11 Oct 2026
- Storage
- M3DB is a distributed time-series database for storing real-time metrics at long retention.m3db.io · read 11 Oct 2026
- Support
- The project links to a community Slack and GitHub repository.m3db.io · read 11 Oct 2026
