Axiom

What it does

Axiom stores logs, traces, metrics, and events on a managed event store and lets you query them with APL, a piped query language built for log-style workflows (the homepage shows queries like filtering status >= 500 and summarizing p95 latency by route). A second language, MPL, covers metrics on a separate purpose-built engine.

The architecture is schema-on-read: there are no indexes to manage and no ingest-time schema decisions. Virtual fields handle transformation at query time instead.

Features visible on the site

• Query, Stream, Dashboards, Monitors, Datasets — the core product surfaces shown in the interface demo
• Broad ingest routes — the datasets shown include OpenTelemetry collector data, Vector, Fluent Bit, Kinesis Firehose, Lambda events, GitHub push and pull-request events, and Vercel web analytics
• AI-assisted investigation — an axiom-agent connected via Axiom MCP answers questions like "Why did checkout p95 spike at 14:21 UTC?" by querying metrics, narrowing the incident window, and correlating traces, logs, and release events
• Dataset sharing — some datasets are marked SHARED, e.g. OpenTelemetry demo traces and Vercel web analytics

Pricing

Axiom prices on a single usage-based dial across logs, traces, metrics, and events, with automatic volume discounts in-console and per-unit rates that drop at higher tiers. The homepage states there is a permanent free tier and self-serve enterprise add-ons.

The homepage's pricing calculator shows an estimated $25/month at 1 TB of ingest and claims roughly 95% savings versus the cheapest alternative, with specific comparisons: Splunk Observability Cloud (137.2x more expensive), Datadog Logs + Metrics + APM (111.4x), Grafana Cloud Pro (20x), and self-hosted ELK + S3 on GCP/AWS (20x). These are Axiom's own claims, not independent benchmarks. Exact current rates are on their pricing page.

Who uses it

The homepage quotes Bailey Pumfleet, co-founder and CEO of Cal.com, and a team at Claras.ai. Logos shown include Asana, Bedrock Robotics, Carbon3D, Convex, Eventim, Factory.ai, Framer, and Luma Labs.

FAQ

What query language does Axiom use? APL for logs and events — piped and sequential, in the style Splunk users know — and MPL for metrics.

Does Axiom support OpenTelemetry? Yes. The homepage shows an otel-collector dataset collecting OpenTelemetry logs, metrics, and traces, plus a shared dataset of OpenTelemetry demo traces.

Is there a free tier? Yes, the homepage describes it as a permanent free tier.

How much does Axiom cost? Pricing is usage-based. The homepage calculator estimates $25/month for 1 TB of ingest, but per-unit rates and tier details aren't spelled out on the homepage — check axiom.co/pricing for current numbers.

Do I need to manage indexes or schemas? No. Ingest is schema-less and schema-on-read, with virtual fields for query-time transformation.

Alternatives

• Datadog — Full-stack observability: metrics, traces, logs, RUM.
• Sentry — Error tracking and performance monitoring for every stack
• Highlight — Open-source full-stack monitoring: errors, replay, logs.
• Better Stack — Uptime monitoring, incidents, and status pages.