Do it Yourself: Generic REST API-Based Monitoring

DX Unified Infrastructure Management (DX UIM) has more than 150 monitoring probes, which enables IT administrators to monitor everything from traditional mainframe servers to modern hybrid clouds running on a wide range of platforms and operating systems.

Traditionally, a separate probe has been required to monitor each specific technology. That’s because the interface that retrieved monitoring metrics was either proprietary or technology specific.

For example, in the context of storage monitoring, DX UIM has more than 18 different probes in its portfolio, enabling comprehensive coverage of storage environments.

In the past, adding monitoring coverage of newer technologies had been a significant project that spanned multiple agile sprints and required a significant investment of staff hours. The average turnaround time for adding monitoring support for any new technology was four to six weeks. This lengthy delay represented a problem that needed to be solved.
As the industry matured and client-server communication became API driven, infrastructure vendors also started exposing APIs for monitoring data and alarms. One of the most popular types of API is REST or, as they’re sometimes known, RESTful APIs. (REST stands for Representational State Transfer.) REST APIs were designed to take advantage of existing protocols. While REST can be used over nearly any protocol, when used for web APIs it typically takes advantage of HTTP. One of the key advantages of REST APIs is that they provide a great deal of flexibility. Data is not tied to resources or methods, so REST can handle multiple types of calls, return different data formats, and even change structurally with the correct implementation of hypermedia.

With the goal of reducing the release time for the support of newer technology within DX UIM, we did some internal experiments to develop monitoring probes that leverage REST APIs for monitoring data collection. The first probe released based on this approach was a monitoring solution for a popular storage technology, XtremIO. This effort took us just two weeks, from inception to release. Leveraging the efficiency of this REST-based monitoring approach, we then went on to quickly deliver custom solutions for customers. For example, we developed support for point-of-sale servers used by retailers, proprietary devices for the entertainment industry, and custom video conferencing systems. As we built these solutions, our framework matured and we ultimately reduced our turnaround time for custom probes to less than a week.

With its robust architecture and scalability, DX UIM is extremely popular with customers. Following are two key advantages of the solution:

• Breadth of technologies monitored out of the box.
• Ease of extensibility and customizations.

In order to continue to build on these advantages, we decided to package this REST-based monitoring framework and deliver it to our customer base, so they could harness its power. Now, RESTMon is available in DX UIM. RESTMon empowers our users to monitor any technology that’s capable of sending monitoring data via the REST API, from coffee machines to Google Cloud Platform.

RESTMon: How it Works

RESTMon allows the user to create custom probes for any technology they need to monitor. To create a probe, teams can use a simple, wizard-based interface. To get started, tool administrators follow a three-step process:

1. Download the schema template provided. The template then needs to be updated with these types of details:
   • Metrics that need to be monitored
   • URLs that will be used to fetch data via the REST API
   • Formulas for derived metrics
   • Alarm types and thresholds
2. Upload the completed template.
3. Validate and deploy the custom schema.

RESTMon supports more than 25,000 configuration items, which can be set for metrics being configured in the schema.

RESTMon features an intuitive probe configuration workflow, leveraging a modern setup wizard based on newer HTML5 code and an intuitive operator console designed by our user experience team.

For every schema uploaded, a tile is added in the setup wizard. These tiles can be accessed by IT administrators and operators. Using modern, HTML5-based setup screens, they can configure monitoring profiles for the newly added technology.

Once monitoring is started, DX UIM automatically populates performance reports and dashboards in the OpConsole. The screen shots below illustrate the results of monitoring computing in Google Cloud Platform.

Even if the technology being used in the data center is not supported out of the box by DX UIM, RESTMon makes it easy to create a custom probe. This enables users to gain a single-pane-of-glass view for their entire infrastructure. IT operators responsible for monitoring data centers and triaging issues will not see any difference between data collected via these custom probes and any commercially available probe in DX UIM. These operators will be able to have a consistent experience in terms of metric style, dashboards, reports, and even SLA management.

For more information, please visit the DX UIM home page at the Broadcom Enterprise Software Academy.