Elastic Observability

A shared home for operational signals

Elastic Observability is the observability-focused solution built around Elasticsearch and Kibana. Its scope includes log and infrastructure monitoring, application performance monitoring, distributed tracing, real-user and synthetic monitoring, alerting, and SLOs. The value of the combined view is practical: an engineer can inspect a service trace, related logs, and host or container metrics in the same analytics environment instead of treating each signal as a separate tool. It is not a separate general-purpose search engine from Elastic Stack; it is Elastic's named observability solution built on that stack.

Start with one service or host

Choose a deployment first. Elastic documents self-managed Stack, Elastic Cloud Hosted, and fully managed Serverless options. A first application-monitoring task needs an Elastic deployment or Serverless project, an application instrumented with an Elastic OpenTelemetry SDK, and a collector or Elastic Agent to forward telemetry. The documented quickstart also calls for an APM agent key for the Serverless path; the self-managed Stack path uses Elastic Agent to collect and forward the application's OpenTelemetry data. The exact setup differs by deployment.

Once data arrives, confirm traces, metrics, and logs in the Applications view, inspect transactions and spans, and use the service map to follow dependencies. For infrastructure monitoring, onboard host or service data through an Elastic integration or OpenTelemetry. The getting-started guide describes prebuilt dashboards, ES|QL for searching and filtering, and rules that can notify external services. A team can then define an SLO around availability, response time, or error rate. For example, an on-call engineer investigating a slow API could compare trace latency with host metrics and related logs before deciding whether the bottleneck is in the service or its infrastructure.

Deployment and feature trade-offs

Serverless reduces cluster operations because Elastic manages the underlying Elasticsearch and Kibana instances. Hosted deployments provide more direct control over cluster resources, while self-managed deployments leave hardware, orchestration, sizing, and version management to the operator. Elastic's pricing page describes hosted and Serverless usage-based pricing and a separate self-managed license model. Its self-managed subscription page lists a no-cost Basic tier alongside paid tiers, but also notes that feature availability and licensing vary. The Observability Serverless getting-started page is specifically for the Complete tier and points Logs Essentials users to a different guide, so check the current feature matrix before assuming all deployments include the same capabilities.

Application telemetry is not automatic: teams need instrumentation and a collector or agent, and may need to configure authentication and data routing. That gives teams flexibility to use OpenTelemetry but adds setup work. This guide is based on current official documentation; no Elastic account was created and no telemetry was sent.