Performance monitoring is an essential practice in network monitoring. When something goes wrong with a device, be it a physical server, a network storage system, or a virtual switch, there are often signs or symptoms. These symptoms might display in various places, and they could be related to the CPU, to the hardware, or maybe bandwidth usage. Only by tracking them can you be aware of performance issues.
For instance, an inexplicably high CPU utilization for an IBM blade server can be tracked to the inefficient cooling of its chassis unit. An IBM power virtual server suffering from high latency can be caused by zombie VMs causing a virtual sprawl in its host server.
Without proper performance monitoring, you can’t see any of the symptoms of an impending issue. They can snowball into something major and take out your services. In this blog, we’ll tell you how ManageEngine OpManager can help prevent this, referencing IBM performance monitoring as an example.
IBM: A solution for every problem
Why IBM? Its solutions are among the most popular and IBM’s blade servers, power servers, and AIX server software, especially, are used in networks around the world. IBM itself is among the top five vendors in terms of market share in several segments and its market presence dates back more than 100 years. Its among the top five vendors in terms of market share for servers and storage devices.
IBM offers IT product solutions in the following categories:
- Servers
- Virtualization
- Storage devices
- Routers & switches
- Load balancers & printers
- Application infrastructure
- Software as a service
IBM has also been a proponent of research in new technologies, promoting studies in open source solutions, quantum computing, and language models in recent years.
IBM performance monitoring with OpManager
OpManager enhances IBM performance monitoring with its vendor specific performance monitors. OpManager has general performance monitors for all kinds of vendors and device types, as well as specific monitors for specific vendors and device types. The specific monitors are usually more accurate.
But first, let’s take a look at the term “performance monitor”. What does it refer to?
Generally speaking, a performance monitor is something that you can apply to obtain a performance metric from a device. An IBM CPU utilization monitor will provide you with the CPU utilization for an IBM device at a specified interval, like one minute.
Performance monitors have a unique vendor specified Object Identifier (OID). OpManager uses these OIDs to gather these metrics and show them to you. Protocols like SNMP and WMI are used to get this information. Configuring these OIDs and monitors sound like tough work, eh? OpManager features over 10,000 device templates with OIDs specified for each device type ready out-of-the-box, including more than 50 IBM device templates for IBM performance monitoring.
IBM performance metrics you can monitor with OpManager
OpManager features an extensive variety of IBM performance monitor types. Here are some of them:
Traffic monitoring: OpManager can monitor incoming traffic, outgoing traffic, and traffic utilization per interface for IBM devices. It enables you to also monitor network sessions like the number of TCP ports in listen state which helps detect abnormal traffic patterns, and plan bandwidth utilization to prevent over use.
Hardware monitoring: Hardware health monitoring helps you prevent unforeseen pitfalls that can occur to your devices. This is particularly important for high density assemblies like server and storage racks that generate a large amount of heat when operating. OpManager can monitor hardware metrics like the temperature of various device components, fan speed in rpm, chassis health status, and the voltage of the power supply units.
OpManager also supports uninterruptible power supply (UPS) monitoring for UPS devices that provide power redundancies to server assemblies.
CPU performance monitoring: Ensuring good CPU health is paramount and you need visibility into different CPU cores and components to accomplish this. With OpManager, you can monitor CPU performance indicators like CPU utilization, memory utilization, processor clock speed, memory data width, and CPU temperature. Real-time data can be obtained with intervals as low as 10 seconds for proactive monitoring.
General health indicators: In addition to CPU, hardware, and traffic indicators, OpManager can also help monitor IBM performance using health monitors like self-test failures and successes, failed maintenances, total uptime, and time elapsed since the last maintenance. You can set alerts for these monitors so you will be notified when your IBM device is in bad health.
IBM blade monitoring: Monitor the performance of your IBM blade servers by keeping tabs on system health status, power state, temperature, blower speed, health state of each module, and other relevant metrics.
Storage performance monitors: OpManager also helps you track IBM storage devices, including IBM flash modules, RAIDs, tape libraries, and more. You can use OpManager to monitor the health of storage devices, storage usage, and forecast storage for capacity planning.
Alerts to complement IBM performance monitoring
Prompt alerting enables you to get the best out of performance monitors. Alerts should have three characteristics: First, they should convey information about the situation. This helps you start fixing the issue right away. Second, alert floods and false positives should be avoided as they get in the way of detecting actual issues. And third, alerts should provide options that can be followed up with actions.
Let’s see if OpManager can satisfy all three of these conditions.
First, alerts can be color coded with five varying levels of severity: attention, trouble, critical, service down, and clear. In addition, OpManager provides basic information about the nature of an issue when it generates alerts.
To prevent alert floods and false positives, you have the option to enable adaptive thresholds. Adaptive thresholds are set using three days of data from the network. On days of low network activity, the threshold itself will be low, and vice versa.
You can also perform actions on alarms such as setting an alarm escalation profile, triggering an automated workflow, managing the device remotely, running root cause analysis, and so on.
Whatever is wrong with your IBM device’s performance, OpManager will monitor it, detect it, and alert you.
How complete governance with OpManager eliminates IBM performance issues
With OpManager, you can monitor the performance of your IBM infrastructure and get alerted about any discrepancies. By proactively detecting and addressing issues, you can avoid drops in quality, outages and other unwanted scenarios.
Let’s review an example. Say, you have an important service hosted on an IBM server rack. The cooling system has a separate power unit from the rest of the assembly, a power outage occurs and your power backup takes over, but the power backup for the cooling system fails. Using a normal performance monitoring tool, you’ll only know about some of these issues when your clients complain about the service going down.
But with OpManager, you’ll receive alerts about increasing server temperature and the cooling power supply going down—proactively. Before your service goes down and your customers are affected, you can fix the issue.
Whether its IBM performance monitoring or any other device or vendor, OpManager has got you covered. Still having doubts? Why don’t you see for yourself with our free, 30- day trial? You can also schedule a free, personalized demo to see how OpManager fits in your network.