Aodh is the OpenStack telemetry alarming service that evaluates metrics against user-defined alarm rules and triggers notifications when conditions are met (observability / monitoring).

Alarming service for OpenStack Telemetry that evaluates metric data against thresholds (observability / monitoring).
Supports multiple alarm types, including threshold-based and combination alarms on telemetry data (observability / monitoring).
Integrates with Ceilometer and other telemetry data sources in OpenStack for alarm evaluation (cloud infrastructure / integration).
Dispatches alarm notifications through pluggable notifiers, such as webhooks and message queues (eventing / notification).
Provides Representational State Transfer (REST) APIs for alarm management, configuration, and retrieval by other OpenStack services or external tools (API / integration).

More About Aodh

Aodh is the alarming component within the OpenStack Telemetry framework and provides alarm evaluation and notification based on metrics and events produced by other OpenStack services (observability / monitoring). It is designed to separate alarm processing from metric collection, which is typically handled by Ceilometer or other telemetry backends. By focusing on alarm rules, state transitions, and notifications, Aodh enables operators and automation systems to react to metric trends and threshold breaches in OpenStack environments.

The core purpose of Aodh is to let users define alarm conditions on metered data, such as resource utilization or service metrics, and to manage the lifecycle of those alarms. It exposes a REST Application Programming Interface (API) through which alarms can be created, updated, listed, and deleted (API / integration). Alarm definitions include parameters such as the metric name, statistic, comparison operator, threshold, evaluation period, and action endpoints. Aodh periodically evaluates incoming metrics from supported data sources against these rules and updates the alarm state, typically between OK, ALARM, and INSUFFICIENT_DATA, based on configured criteria.

Aodh supports several alarm types, including threshold alarms that compare a single metric to a defined value, and combination alarms that combine the states of multiple underlying alarms (observability / monitoring). This allows operators to express composite conditions that consider multiple resources or services. The service also supports event-based alarms when integrated with telemetry event streams, enabling alarms to trigger on discrete events instead of or in addition to time-series metrics.

Notification delivery in Aodh is handled through a pluggable notifier system (eventing / notification). Alarm actions can be configured to send callbacks via webhooks, publish messages to queues, or trigger other OpenStack workflows, such as autoscaling through orchestration services. This makes Aodh a central alarm engine that other platform components can use to initiate remediation, scaling, or ticketing processes.

In enterprise and institutional deployments, Aodh typically runs as part of a broader OpenStack cloud, working with services like Ceilometer for metering and Gnocchi or other backends for metric storage (cloud infrastructure / integration). Operators use Aodh to monitor compute, storage, and network resources, and to integrate alarm signals into existing operational tooling. The architecture usually includes Aodh API, evaluator, notifier, and listener processes, backed by a database and message bus consistent with OpenStack patterns (cloud infrastructure / architecture).

From a categorization perspective, Aodh fits into telemetry alarming and notification within cloud infrastructure monitoring. It provides an API-driven, multi-tenant alarm management layer that aligns with OpenStack’s service-oriented design, enabling other services and external systems to consume alarm states and notifications for automated operations and incident handling.