Skip to main content

avuru obs vs. SigNoz

SigNoz is the closest tool to avuru obs on paper: both are OpenTelemetry-native and both store every signal in ClickHouse. It's a strong, actively developed platform. The differences are three: how services get onto the map (eBPF vs. SDK instrumentation), whether continuous profiling is a first-class signal, and where the open-source line is drawn.

At a glance

SigNozavuru obs
Primary scopeTraces, logs, metrics, APMTraces, logs, metrics, profiling
InstrumentationOpenTelemetry SDKs (manual)eBPF auto-discovery (OBI) + optional OTLP spans
StorageClickHouseClickHouse
Service mapFrom instrumented spansDerived from spans + eBPF — even un-instrumented services appear
Continuous profilingNot a core signalBuilt in (opt-in) as the 4th signal
IngestNative OTLPNative OTLP (:4318/:4317)
DeployMultiple componentsOne Helm chart
LicenseMIT core; SSO, RBAC & more gated behind commercial EnterpriseAGPL-3.0 — nothing gated

Why teams look at avuru obs

  • A map without instrumenting first. SigNoz shows what your SDKs emit; avuru's sensor runs upstream OBI to reconstruct topology from the kernel, so un-instrumented services still appear. See Architecture.
  • Profiling in the box. Continuous CPU profiling is the fourth signal in the same ClickHouse store, correlated with traces — not a separate tool to add.
  • No open-core line. SSO/OIDC, RBAC and similar are avuru features, not an Enterprise upsell. Everything ships under AGPL-3.0.

How to migrate

You're already on OpenTelemetry, so this is a one-line change:

  1. Install avuru obs alongside SigNoz.
  2. Point your OTLP exporter (or OTel Collector) at the avuru gateway on :4318/:4317. Dual-export to both while you validate.
  3. Confirm the data matches, then retire the SigNoz pipeline.

See the OTLP bridge guide for exporter settings.

When SigNoz is the better fit

  • You want a permissive MIT core rather than avuru's copyleft AGPL-3.0.
  • You rely on its managed Cloud, its dashboards, or its broader APM feature catalog today.
  • You don't run on Linux/Kubernetes, where the eBPF map adds the most value.

:::tip Try it side by side Install in 30 seconds and dual-export one service — the eBPF map appears next to the traces you already send SigNoz. :::