Observability Overview¶

The StarlingX OpenStack application provides observability capabilities that give you visibility into the health, performance, and operational state of your cloud. With components such as the Ceilometer exporters, OpenStack Exporter, Aodh services, and the optional prometheus platform application, you can collect and store metrics, process service events, evaluate alarms, and expose telemetry data to monitoring and analytics systems.

Key Components of the Observability Stack ¶

The following table serves as a reference for the major observability components and their operational characteristics.

Component	Chart	Namespace	Purpose	Consumed By	Default	How to Change
Ceilometer Central Exporter	`ceilometer`	`openstack`	Collects metrics from OpenStack services.	Prometheus server and external monitoring systems.	Enabled	Modify the Ceilometer Helm overrides and reapply the application.
Ceilometer Compute Exporter	`ceilometer`	`openstack`	Collects metrics from compute hosts and hypervisors.	Prometheus server and external monitoring systems.	Enabled	Modify the Ceilometer Helm overrides and reapply the application.
Ceilometer Notification Agent	`ceilometer`	`openstack`	Processes OpenStack notifications and generates events for alarm evaluation.	Aodh event-based alarms.	Enabled	Modify the Ceilometer Helm overrides and reapply the application.
OpenStack Exporter	`prometheus-openstack-exporter`	`openstack`	Exposes cloud-wide metrics collected from OpenStack APIs.	External monitoring systems.	Disabled	Enable the exporter through the Helm chart overrides and reapply the application.
Prometheus platform application	`kube-prometheus-stack`	`monitoring`	Collects and stores metrics from the Ceilometer exporters and provides query capabilities for metric-based alarms.	Aodh alarm evaluator and external monitoring tools.	Uploaded, not applied	Apply the Prometheus application.
Aodh services (API, evaluator, listener, notifier)	`aodh`	`openstack`	Evaluates metric-based and event-based alarm conditions and executes alarm actions.	Operators and external notification endpoints.	Enabled	Modify the Aodh Helm overrides and reapply the application.

Architecture ¶

The observability stack consists of the following frameworks:

Metrics Framework: Collects and stores metrics from OpenStack services to support monitoring, visualization, and metric-based alarm evaluation. For more information, see Metrics Framework.
Event Alarm Framework: Processes OpenStack service notifications and evaluates event-based alarms. Events flow from OpenStack service notifications through RabbitMQ and the Ceilometer Notification Agent, and are then evaluated by the aodh-listener service for event-based alarms. For more information, see Event Alarm Framework.

These frameworks operate independently. If you only require event-based alarm functionality, you do not need to deploy the prometheus application. Similarly, metric-based alarms can be configured and evaluated without using OpenStack service notifications or events.

The common integration point between the two frameworks is the Aodh API, which is used to create, manage, and query alarm definitions.

Metrics Framework ¶

The Metrics Framework includes the following exporters, which collect and expose metrics in Prometheus format:

Ceilometer Central Exporter collects metrics by polling OpenStack service APIs, such as Nova, Neutron, Glance, and Cinder. These metrics provide visibility into the state and resource usage of OpenStack services. Metrics are exposed on port 9101 through the Service osh-openstack-ceilometer-central.
Ceilometer Compute Exporter collects metrics from the local hypervisor on each compute host, providing visibility into compute node and instance resource usage. Metrics are exposed on port 9101 through the Service osh-openstack-ceilometer-compute. This service is configured with internalTrafficPolicy: Local and responds only on the node where the corresponding exporter is running.
OpenStack Exporter collects inventory and status information from OpenStack services, including resources such as instances, hypervisors, networks, ports, images, volumes, and stacks. It polls the OpenStack APIs at a configurable polling interval (default: 30 seconds) and publishes metrics in Prometheus format intended for external monitoring or analytics systems (not collected by the prometheus platform application). Metrics are exposed on port 9180 through the Service openstack-metrics.

The prometheus platform application discovers these exporters through ServiceMonitor resources, collects their metrics on a fixed interval, and stores the metrics in its time-series database (TSDB). The aodh-evaluator service then queries the stored metrics to evaluate metric-based alarms.

For more information, see Configure OpenStack Exporter for Metrics Collection.

Note

Metric-based alarms require the prometheus platform application. For details on the exporters, ServiceMonitors, and collection intervals, see Configure Ceilometer for Telemetry Data Collection.

Event Alarm Framework ¶

The Event Alarm Framework enables alarms to be triggered by OpenStack service events.

OpenStack services publish notifications to RabbitMQ when operational events occur. The Ceilometer Notification Agent converts these notifications into Ceilometer events, and the aodh-listener service evaluates the events against event-based alarms. When an alarm changes state, the aodh-notifier service sends a notification to the configured alarm action endpoint over HTTP or HTTPS.

Because this path does not involve the Prometheus server, event-based alarms work even when the prometheus platform application is not applied.

Note

For details on creating event-based alarms, configuring notification endpoints, and event trait filtering, see Manage Alarms Using Aodh.

Observability Key Workflows and Tasks ¶

The following table provides an overview of the key observability workflows and the corresponding sections in this document.

Step	Area	Description	Section
1	Telemetry Collection (Ceilometer)	Configure the Ceilometer central and compute agents to collect and expose metrics from OpenStack services and compute hosts.	Configure Ceilometer for Telemetry Data Collection
2	Metrics Storage (Prometheus)	Deploy and configure the `prometheus` platform application to scrape, store, and query metrics collected by the Ceilometer exporters.	Configure Prometheus for Metrics Storage
3	Alarm Management (Aodh)	Create and manage metric-based and event-based alarms. Aodh evaluates alarm conditions and triggers notifications to configured endpoints.	Manage Alarms Using Aodh
4	External Integration	Expose telemetry endpoints and forward alarm notifications to external monitoring or analytics systems, including the OpenStack Exporter.	Integrate with External Systems