|Visualization of data sounds like an easy proposition; collect data and display it. But there’s a lot more to it than that! What data should be displayed? If you collect data from, say, 200 systems, do you want to see data about each system or an overview of them all? If half of them are Windows systems and half are Linux systems, you might want to see overall utilization of each group. Or, how about a view that shows which systems might be underutilized in terms of resources used?|
|We call this discussion the Nerd View vs. the Management View. Just ask our development team!|
|The classic nerd here at Sightline likes data, and lots of it. We want to see the details about every system, its resources, its workloads, throughput, you name it. We need drill-down to the process level data, and the ability to set alerts. Basically, if it can be collected and displayed, we want it.|
|Management requirements are a little different. Management needs to know that the environment is in good shape, but still needs to retain the capability to drill down if necessary. Management wants to know that the critical applications are up and running and making money for the company.|
|Using EDM, you can have your cake and eat it too!|
|Consider this display, a typical nerd view. Lots of detail about the system workload: cpu and memory breakdown per workload, and the numbers of threads and processes for each workload. Event data would fall into this category, as well.|
|On the other hand, management wants a higher-level view, or a view to support a specific application, service or report. They want to see something like this: an overall view of resource consumption, as in Sightline’s new Utilization Reports. This shows which systems are over- or under-utilized in terms of cpu, memory, disk and network resources.|
|With the depth of data available from Sightline data collection agents, we can show whatever you need to see. Remember, we’ve got IT covered.|
This week’s Ask John question is from Oscar Jiminez at Megadata Information Solutions in Costa Rica. Thanks, Oscar!
Question: In an EDM high availability hypothetical scenario (2 EDMs working as active/active nodes), can both EDMs can share the same EDM DB? Under this scenario, I can put this database in a SQL cluster and provide it also high availability.
Answer: Certainly. This implementation is supported; click here for an example of configuring JBoss with Apache HTTPD.
The Sightline Power Agent for Linux Systems is required to be installed from a user with full root access. However, this is not always possible. This document provides steps to grant root privileges to non-root users, so that they can install and run the Sightline Power Agent on Linux systems.
Linux uses RAM memory for disk caching to speed up the system. This means that the Mem % Free metric will consistently be low, perhaps 5%, when the system is actually only using 50% of the RAM. In Sightline, this real usage can be monitored and alerts can be created to alert you if the amount of real memory goes too low.
In addition to the in-depth operating system metrics that Sightline collects for virtually all platforms, Sightline can collect and analyze performance metrics for many Java J2EE servers including IBM WebSphere Application Server and RedHat JBoss. This allows alerting, analysis and reporting of all the critical metrics related to EJBs, JMS Queues, Connection Pools, memory and threads. Find issues faster and track them down using the advanced analytics found only in Sightline. Click here for details.
Q: Can I display the sum of a metric across multiple data sources?
A: Yes. Visualization in EDM includes a sum across data sources feature. When you create a chart, click the Sum Across Data Sources checkbox. Add the chart to a template, then create a report using that template and assign multiple data sources to the report. When the report is displayed, the target metric(s) will be summed for all of the data sources selected in which it appears.