What it does
OpenSearch Observability Stack is a preconfigured observability deployment assembled from OpenTelemetry Collector, Data Prepper, OpenSearch, Prometheus, and OpenSearch Dashboards. It accepts logs, metrics, and traces, and includes views for service maps, application performance, and AI-agent traces. Teams can start with Docker Compose and included example services, or deploy the same group of components to Kubernetes using Helm. It is useful for engineers who want to explore an integrated observability stack locally or operate its components in their own environment.
Explore sample telemetry before instrumenting an app
The local installation guide requires Docker and Docker Compose, macOS, Linux, or WSL2 on Windows, and at least 8 GB of available memory. Clone the official repository and start its Compose stack, then open OpenSearch Dashboards at the documented local address. Example services can generate telemetry immediately, so a first exploration can look at pre-populated traces, logs, metrics, and service maps without changing application code. For your own service, the documentation describes sending OpenTelemetry data or using its no-code eBPF instrumentation path.
The project also publishes a Helm-based Kubernetes deployment path. That route installs the same core services and therefore moves the exercise from a laptop demo into cluster operations, persistent storage, credentials, and component sizing. This is a self-managed stack, not a hosted incident-response service. The repository lists Apache License 2.0, while the sources reviewed here did not establish a separately priced managed edition.
Treat the included credentials as local-development defaults only. The installation guide instructs operators to change the default password before using a non-local deployment. These setup instructions were reviewed but the stack was not installed or run.
Strengths
- Packages an OpenTelemetry Collector, Data Prepper, OpenSearch, Prometheus, and dashboards into one documented deployment.
- Includes sample services and workflows for traces, logs, metrics, service maps, and AI-agent telemetry.
- Can be started locally with Compose or deployed to Kubernetes with Helm under an Apache 2.0 license.
Watch out for
- The local setup requires Docker Compose and at least 8 GB of available memory; a Kubernetes deployment requires Helm and a configured cluster.
- Operators remain responsible for storage, credentials, updates, and the health of several self-managed components.
Best for
- Developers exploring a complete OpenTelemetry observability stack with sample data
- Platform teams that want to self-host logs, metrics, traces, and service maps
- Teams instrumenting AI-agent execution with OpenTelemetry conventions
Sources
- OpenSearch Observability Stack | OpenSearch - Observability Stack (Oct 26, 2026)
- Installation | OpenSearch - Observability Stack (Oct 26, 2026)
- opensearch-project/observability-stack (Oct 26, 2026)
- Apache License 2.0 - OpenSearch Observability Stack (Oct 26, 2026)