Skip to main content
OpenTelemetry is an open standard for collecting traces, metrics, and logs from your services. Obsy wraps the full OTel lifecycle — from collector deployment to telemetry quality reviews.

​
What Obsy deploys

For each Kubernetes cluster you register, Obsy installs:
ComponentTypeRole
OTel GatewayDeployment (1+ replicas)Receives OTLP from your services; applies sampling, filtering, and batching; exports to Datadog/Grafana
Node CollectorDaemonSet (one per node)Collects host metrics, kubelet stats, and container logs from every node
Both are deployed to the obsy-system namespace via the official open-telemetry/opentelemetry-collector Helm chart.

​
Pipeline

Your service (OTel SDK)
        ↓  OTLP gRPC/HTTP :4317/:4318
   OTel Gateway (obsy-system)
        ↓
  ┌── Memory Limiter ──────────┐ drops data if memory > soft limit
  ├── K8s Attributes ──────────┤ enriches with pod/namespace/node
  ├── Health Check Filter ─────┤ drops /health, /ping, /metrics
  ├── Custom Drop Rules ───────┤ optional: drop by endpoint, method, user-agent
  ├── Sampling ───────────────┤ probabilistic or tail-based
  └── Batch Processor ─────────┘ buffers before export
        ↓
  Datadog / Grafana / Both

​
OTel features in Obsy

Collector management

Install, monitor, and reinstall collectors per cluster.

Sampling

Reduce telemetry volume and cost without losing critical signals.

Telemetry Review

Per-service quality and cost analysis with AI-powered recommendations.

Reliability Reports

OTel collector health reports with golden signal coverage.

Health Monitors

11 pre-built monitors for collector health in Datadog or Grafana.

​
Telemetry signals

Obsy works with all three OTel signal types:
SignalWhat it capturesPrimary use
TracesRequest spans across servicesLatency analysis, RCA, dependency mapping
MetricsCounters, gauges, histogramsDashboards, monitors, SLOs
LogsStructured and unstructured log recordsCorrelation, debugging