OpenTelemetry (OTel) is an open-source framework for collecting telemetry data—traces, metrics, and logs—with the goal of standardising observability across distributed systems. Backed by the Cloud Native Computing Foundation (CNCF), OTel is rapidly becoming the industry standard for instrumentation in modern architectures.
Why OTel Matters
The primary advantage of OTel is its vendor neutral approach. Instead of locking into a proprietary agent or SDK, OTel provides a consistent model for instrumentation and data export. This gives organisations full control over how they implement and operationalise observability, enabling flexibility in backend choices, whether that's Prometheus, Jaeger, Honeycomb, or a managed APM solution.
OTel supports two main instrumentation strategies:
Auto-Instrumentation (Zero-Code)
For common languages and frameworks, OTel provides agents that automatically capture telemetry without requiring code changes. This is ideal for teams looking for quick wins with minimal development effort.
Manual Instrumentation
For complex or custom systems, OTel exposes APIs and SDKs that allow developers to instrument code directly. While powerful, this approach demands additional development and testing, making it more suitable for organisations with mature engineering practices.
The Trade-Off
While OTel offers flexibility and cost control, it introduces operational overhead. Managing collectors, exporters, and pipelines requires planning and governance. Compared to platform based observability solutions, OTel can feel heavy for organisations without strong CI/CD workflows or centralised DevOps capabilities.
Managed Observability Platforms
Solutions like Dynatrace, Datadog, and New Relic have invested heavily in out-of-the-box, agent-based instrumentation. These platforms minimise complexity by providing turnkey deployment and integrated analytics. For many use cases, installing a vendor agent and configuring server-side settings is significantly easier than building and maintaining an OTel pipeline.
While these platforms can ingest OTel data, their core value proposition often lies in proprietary instrumentation and advanced features—such as AI-driven anomaly detection and automated root cause analysis.
OTel-Oriented Platforms
Platforms like Honeycomb, Elastic, and the Grafana stack are designed with OTel in mind. They offer strong visualisation and correlation for OTel data and often integrate directly with OTel collectors, reducing setup complexity. For organisations committed to open standards, these solutions provide a natural fit.
When to Choose What
Use an OTel-oriented solution (e.g., Grafana, Honeycomb) if:
- Operational overhead is manageable and CI/CD pipelines are mature.
- Your architecture primarily uses OTel supported languages (Java, .NET, Go, Python, etc.).
- You prioritise vendor neutrality, cost control, and flexibility.
Use a managed observability platform (e.g., Dynatrace, Datadog) if:
- Your organisation is large, with a mix of legacy and modern systems.
- Operational overhead is a concern (limited DevOps or testing resources).
- Teams are siloed and need fast, low touch observability.