Oracle RAC monitoring
Site24x7 supports monitoring of Oracle Real Application Clusters (RAC)—auto-discovering your RAC cluster once the agent is installed and Oracle Database credentials are provided. Track cluster-wide performance indicators, per-instance metrics, and per-node Clusterware health from a unified dashboard.
How Oracle RAC monitoring works
In an Oracle RAC environment, multiple instances run across different nodes in a cluster. Site24x7 collects performance and health metrics by connecting directly to these instances.
Connection behavior
When configuring Oracle monitoring, the hostname provided determines how the agent connects to the database.
If a SCAN listener hostname (for example, RACCLUSTER-SCAN) is used, it resolves to multiple node IP addresses in a round-robin manner.
Using direct node hostnames ensures the agent consistently connects to the same instance. This delivers stable data collection and accurate tracking of instance‑level metrics, helping teams gain reliable insights without interruptions.
How to create an Oracle RAC monitor
For consistent and accurate monitoring of individual Oracle RAC instances, deploy a Site24x7 agent on each RAC node and configure a dedicated monitor per node using that server's hostname along with the appropriate instance-specific credentials.
Example configuration
To monitor an Oracle instance running on a specific node, set the hostname to the node's fully qualified hostname:
Hostname: dbnode01.example.com
With this setup, the Site24x7 agent consistently connects to the same Oracle instance, enabling an accurate collection of per-instance metrics such as system global area (SGA), program global area (PGA) memory usage, and session details.
Prerequisites
Before setting up Oracle RAC monitoring, make sure the following conditions are satisfied:
- Site24x7 Linux agent version 22.2.0 or above.
- Oracle Database version 11g or above.
- Oracle Grid Infrastructure (Clusterware) must be installed and running on the node where the agent is installed. The Clusterware utilities crsctl and olsnodes must be accessible to the agent user.
Provide the required permissions and follow the steps in the Oracle Database monitor help document to add your Oracle Database monitor. After adding the monitor, the Oracle RAC cluster running on the server will be auto-discovered and added for monitoring.
Oracle RAC monitoring is supported on Linux only. The Site24x7 agent must be installed on all RAC nodes, and local listener endpoints must be provided for each node.
Oracle RAC monitor
Once the agent is successfully installed and credentials are provided for Oracle database monitoring, the Oracle RAC cluster running on the server will be auto-discovered and added for monitoring.
To view your Oracle RAC monitor:
- Log in to Site24x7 and go to Database > Oracle RAC in the left navigation pane.
- You will see the list of Oracle RAC clusters that have been auto-discovered and added for monitoring.
- Click a cluster name, then click Instances to view the individual Oracle instances running across all cluster nodes.
If the agent has been upgraded to version 22.2.0 but the Oracle RAC cluster has not been detected, trigger application discovery manually from the respective server monitor. Go to Server Monitor > [specific server] > click the hamburger icon beside the display name > click Discover Applications.
Performance metrics
Monitoring the key performance metrics of your Oracle RAC cluster and its individual instances is critical for identifying bottlenecks and optimizing the overall efficiency of your Oracle RAC environment. View the full list of metrics collected for Oracle RAC monitoring.
Metrics are collected at three levels:
| Level | What is monitored | Data source |
| Cluster level | Aggregated performance indicators, Global Cache Service (GCS) block contention, lock mode distribution, SGA, PGA, undo segment statistics, and cluster-wide wait events. | Oracle GV$ global views |
| Instance level | Per-instance status, configuration details, block contention, and wait events, broken down by instance ID. | Oracle GV$ views filtered by INST_ID |
| Node level | Per-node Oracle Clusterware component health (Cluster Ready Services i.e., CRS, Cluster Synchronization Services i.e., CSS, Event Manager), node number, and node type. | crsctl and olsnodes |
In addition, Oracle Clusterware resource states—including database resources, instance resources, listeners, and VIPs—are tracked for each cluster node using crsctl stat res -t.