Architecture overview

Data collection

Agents and data collectors monitor systems, subsystems, or applications and collect data. An agent or a data collector interacts with a single resource (for example, a system or application) and, in most cases, is on the same computer or virtual machine where the system or application is running. For example, the Linux® OS agent collects performance indicators for the operating system on the Linux host and the WebSphere® Applications agent monitors the performance indicators of WebSphere application servers. Also, some agents track transactions between different resources.

You can set up thresholds on key performance indicators (KPIs). If an indicator changes to go over or under the threshold, the agent or data collector generates an alert, which is the server processes. You can also configure forwarding of events to a target, such as the Netcool/OMNIbus Probe for Tivoli® EIF, Cloud Event Management, or an SMTP server.

You can use Alert Notification to configure email notifications for events.

Agents and data collectors are preconfigured to communicate with the Cloud APM server.

Communication between the server and agents or data collectors

The agents and data collectors on every monitored host establish HTTPS communication with the Cloud APM server, which is in the IBM cloud. The agent or data collector is the client side of the connection.

Agents and data collectors require internet connectivity to send data to the server and, if they cannot send data directly over the internet, a forward proxy might be required. For more information, see Network connectivity.

Data stored by the server

Agents and data collectors push data to the Cloud APM server at intervals ranging from 1 minute to 8 minutes, depending on the type of data. The server stores all values that are sent by the agents and data collectors for 8 days by default. Summarized transaction data is stored for longer periods.

Saved monitoring data is called historical data. The server uses historical data to display tables and graphs that you can use to analyze the trends in your environment.

Historical reports are also available for certain agents. For more information, see Reports.

Scalability

You can monitor up to 10,000 managed systems from Cloud APM. A managed system is a single operating system, subsystem, or application in your enterprise that an agent is monitoring.

Cloud APM supports between 150 and 400 monitored user transactions per second.

Integration

IBM Cloud Application Performance Management integrates with other products and components when they are configured for communication with the Cloud APM server.

Products that can be integrated include IBM Control Desk, Netcool/OMNIbus, Tivoli Monitoring, OMEGAMON®, Operations Analytics - Log Analysis, Operations Analytics - Predictive Insights, IBM Alert Notification, and IBM Cloud.

Agent Builder is a component that can be used to create custom agents.