Observ

Simple pricing

Pay per cluster. Never per volume.

No observability tax. Your bill is decoupled from telemetry volume, retention and seats — they all live on your object storage. You pay for a cluster, not for your traffic.

Community

Free. Unlimited. Self-hosted.

Free

forever — it's your bucket, your volume, your retention

  • Full core: OTLP, DataFusion engine, embedded UI
  • 1 cluster
  • Unlimited telemetry volume & retention (your object storage)
  • OpenTelemetry OTLP ingestion (gRPC & HTTP)
  • Single multi-role Rust binary
  • Unlimited UI seats
  • Community support
Most popular

Pro

Production-ready. Billed per cluster — never per volume.

From590 €/mo

per cluster

5 900 €/year, billed annually

  • Everything in Community
  • Multi-node clusters (up to ~20 compute nodes)
  • RBAC & SSO
  • Alerting & SLOs
  • Advanced dashboards
  • Priority updates
  • Standard support (business hours)

Intelligence

The embedded brain: RCA, anomalies and a local LLM.

+1 490 €/mo

add-on / cluster

14 900 €/year, billed annually

  • Automated Root Cause Analysis (fault trees)
  • BERT/ONNX log classification
  • ML anomaly detection
  • Local LLM chat (MistralRs — no external API)
  • Built-in MCP server
  • Runs entirely inside your infrastructure

Enterprise

Multi-cluster federation, air-gap and managed "BYO bucket".

Custom

5-figure/year · per organization

  • Everything in Pro + Intelligence
  • Multi-cluster federation (Arrow Flight)
  • Air-gapped deployments
  • Managed "Bring Your Own Bucket"
  • SSO / SCIM provisioning
  • Audit & compliance, custom retention
  • SLA & dedicated support
  • Assisted deployment

The observability tax, removed

Usage-based SaaS bills per record or per GB — the cost climbs with every log you ship. Observ stays flat: volume never moves the bill.

TelemetryUsage-based SaaSObserv

Homelab — ~10M records/mo

Evaluation, dev, side projects

Free tier limits, then overageFree

Growing — ~1B records/mo

One production environment

Hundreds–thousands €/mo, climbing with volumeFrom 590 €/mo per cluster

At scale — 10B+ records/mo

Multi-environment, long retention

Tens of thousands €/mo — the taxFlat per cluster — volume doesn't count

+ your own object storage (which you control — and where your data stays)

Full feature comparison

CommunityProIntelligenceEnterprise
Capacity (per cluster)
Clusters1Billed per clusterPer cluster add-onUnlimited
Compute nodes / clusterUp to 3Up to ~20Up to ~20Unlimited
Telemetry volumeUnlimitedUnlimitedUnlimitedUnlimited
RetentionUnlimited (your S3)UnlimitedUnlimitedUnlimited + policies
UI seatsUnlimitedUnlimitedUnlimitedUnlimited + SCIM
Core observability
OpenTelemetry OTLP
DataFusion SQL queries
Embedded UI
Advanced dashboards
Alerting & SLOs
Intelligence (the brain)
Automated RCA (fault trees)
BERT/ONNX classification
ML anomaly detection
Local LLM chat (MistralRs)
Built-in MCP server
Security & governance
RBAC
SSO
SCIM provisioning
Audit & compliance
Air-gapped deployment
Scale & sovereignty
Self-hosted (your bucket)
Multi-cluster federation (Arrow Flight)
Managed "BYO bucket"
Support
Community support
Standard support (business hours)
Dedicated support & SLA
Assisted deployment

Unlimited volume. Unlimited retention. Unlimited seats.

They run on your object storage, not ours — so your bill never moves because you ingested more logs or added a teammate. That's the whole point.

Questions? Talk to us

Questions & Answers

Is the Community tier really free — and really unlimited?

Yes, forever. You get the full core (OTLP, DataFusion engine, embedded UI) on 1 cluster, with unlimited telemetry volume and retention — because they live on your object storage, not ours. Community support, no credit card.

What exactly is a “cluster”?

A cluster is one autonomous Observ deployment: a control-plane / catalog plus the compute pods (ingest, query, compact, llm…) attached to it. One environment — production, staging — is one cluster, no matter how many OTel nodes emit or how many records you store.

Do you ever charge by volume, records or seats?

Never. Volume and retention run on your storage, and UI seats are unlimited everywhere. Billing is decoupled from telemetry by design — that’s the whole promise, and it would be incoherent (and trivially bypassable in self-host) to meter it.

What does the Intelligence add-on include?

The embedded brain: automated Root Cause Analysis (fault trees), BERT/ONNX log classification, ML anomaly detection, a local LLM chat (MistralRs — no external API), and a built-in MCP server. It’s billed as an add-on per cluster, on top of Pro.

What does Enterprise add?

Multi-cluster federation over Arrow Flight, air-gapped deployments, a managed “Bring Your Own Bucket” offering, SSO/SCIM, audit & compliance with custom retention, an SLA and dedicated support, plus assisted deployment.

How does billing work?

Flat per cluster, billed via Stripe. Annual billing is recommended and saves two months versus monthly — no usage-based surprises.

How does this compare to usage-based SaaS?

Per-record / per-GB platforms bill for the volume you ship, so the cost climbs as you scale — the “observability tax”. Observ charges a flat fee per cluster: volume never moves your bill, and your data never leaves your infrastructure.

Need something custom? Let's talk