Metrics monitoring built on the Elasticsearch Platform SREs trust

Elastic pairs best-in-class metrics efficiency with the industry's richest log analytics solution. 30x faster queries than competing TSDBs, built on a columnar datastore engineered for high-cardinality workloads that scale without breaking the bank. With native PromQL so you can keep the workflows you love.

Start free trial
See our benchmarks

Meet the columnar metrics engine that's best in its class

The Elasticsearch columnar datastore outpaces others in metrics ingest, storage, and query speed at any scale.

Scale without dropping data

The engineering depth that set the standard for log ingest, storage, and query performance is exactly what we applied to building a better TSDB for high-cardinality metrics. Same team, same rigor, new data type — built to retain every metric at full resolution without the price tag.

  • BEST-IN-CLASS EFFICIENCY

    Faster queries at a fraction of the cost

    Elasticsearch runs queries 25x faster than Prometheus and stores metrics 2.5x leaner — with no cardinality limits. Keep your current ingest architecture, retain more history, and pay less for it than a comparable Prometheus stack.

  • SCHEMA-AGNOSTIC

    One datastore, all formats

    Most backends normalize everything into a single schema. We don't. Whether you send us Prometheus, OpenTelemetry, Beats, or OCSF, Elasticsearch stores each in its native format and queries it as-is. No translation layer, no information loss, no conversion tax.

  • ONE-DAY MIGRATION

    PromQL from day one

    Your existing PromQL queries, dashboards, and alert rules carry over without having to learn a new language. Remote write and OTLP ingest are both supported. Migration is a configuration change, not a month-long project.

  • LOGS + METRICS + TRACES

    Unified investigations — no context switching required

    In a typical observability stack, finding root cause often means navigating multiple query languages and backends. In Elasticsearch, metrics, logs, and traces are all in one place. When an alert fires, the relevant context is already there.

Elasticsearch doesn't scan rows. It reads columns.

Elasticsearch's segment-based storage is columnar by design, delivering sub-second response on millions of time series with vector loading and processing.

  • Query any data at high cardinality

    ES|QL is built to exploit this: a vectorized query engine that processes data in batches and doesn't degrade at high cardinality. Pipe queries across metrics, logs, and traces — with native PromQL support included.

  • Get more from every byte

    With a full set of time-series functions covering rate, delta, percentile, time bucketing, and aggregations, plus doc value skippers and synthetic ID trimming keeping storage lean, you get more analytical depth without the cost to match.

  • Access from anywhere you already work

    Most backends give you one way in. Elasticsearch gives you three: Kibana for dashboards and prebuilt workflows, Elastic AI Agent for chat-led investigations, and purpose-built MCP apps and skills for the AI tools your team already works in.

ELASTICSEARCH 9.4 BENCHMARKS

Engineering that shows up in the numbers

Head-to-head across the three metrics that define a production-grade TSDB: query speed, storage density, and ingest throughput

Dimension Elasticsearch 9.4 Prometheus Mimir ClickHouse
Query speedHigh-cardinality time series Fastest
Baseline		 Up to 30x slower Up to 30x slower Up to 8x slower
Storage densityBytes/sample Best
3.74 B		 ~9.42 B ~3.95 B ~6.8 B
Ingest throughputSamples/second Fastest
428K/s		 402K/s 404K/s ~300K/s
Native PromQLNo adapter required Native ✓ Native ✓ Native Requires adapter
OTel-nativeNo schema conversion OTel-first Via exporters Via exporters Manual mapping

ELASTICSEARCH AS A COLUMNAR METRICS ENGINE

The innovation that made it possible

From storage architecture to query execution, each part of our platform was built with purpose. Here's the engineering that made it real.

Migration tool — tech preview

Migrate from Datadog or Grafana overnight

Automatically convert dashboards and alerting rules from Datadog and Grafana into Elastic, dramatically reducing the cost and complexity of switching platforms.

Ready to switch and save 50% on your Datadog metrics bill?

Start shipping Prometheus metrics to Elastic

The Prometheus Remote Write endpoint requires no extra configuration. Once metrics are flowing, you can query them with ES|QL using the built-in PROMQL function for PromQL compatibility, or write native ES|QL queries to join metrics with logs and traces in the same store.

Turn metrics into action

Monitor your infrastructure at scale. Explore metrics in Discover, build dashboards as code, and let AI-led investigations highlight anomalies, uncover trends, and automate remediation, so you can plan capacity and resolve issues faster.

Start working with Elasticsearch today

Join thousands of engineering teams who moved from Prometheus, Datadog, and Grafana to a faster, smarter, unified observability platform.

Start free trialContact sales