FAQ

FAQ

Answers to the questions that we are frequently asked about Splunk for WebSphere Application Server.

What are the hardware requirements of my JMX collection machine?

We are working on understanding the best hardware requirements for this. We have tested this at customer sites in their test environments with 60-120+ application servers in a cell and used one machine with 1 CPU to collect all performance metrics without any major impact to that machine. In addition to this, given the functionality to change the frequency of data collection and volume of data collection this can also vary.

What performance impact does JMX collection have on my WAS Deployment Managers?

There is minimal impact on your WAS Deployment Managers. You can control how often the API call is made, how many nodes you are collecting data from, and modules you are collecting. We are working on understanding the best hardware requirements for this. We have tested this at customer sites in their test environments with 60-120+ application servers in a cell and used one machine with 1 CPU to collect all performance metrics without any major impact to that machine. In addition to this, given the functionality to change the frequency of data collection and volume of data collection this can also vary.

Install

Splunk does not install on my WebSphere machine. Can I still get data in, and if so how?

We have a utility that includes log data that is stored anywhere to be pulled into the filtering mechanism of the App (cell, node, server dropdowns), making use of a header in each of the SystemOut.log and native_stdout.log and in some cases SystemErr and native_stderr.log, which contains the cell, node, and server name of that particular log. The events do not contain this information, so a .csv file is created with file paths and cell, node, server for those. File paths are the source field for each event. Host information will not necessarily be accurately available. This data must be added either to the directory structure where this data is stored or to the file name and then transforms.conf must be modified to make this work with our dashboards.

Performance data can be pulled remotely from any WebSphere Application Server Network Deployment installation. WebSphere Application Server for zLinux is a separate product and we have not tested our java calls with that API.

Getting Data in

My log data is not stored in the WAS directory structure but on a different volume, how do I get it in?

If log data is stored in the WAS directory structure but in a different volume, as long as the host data is accurate you can use the provided Splunk WAS specific configuration files to add the second volume so that the script that creates inputs.conf traverses the second volume. There is no limit to the number of volumes that can be added to that configuration file. Note that this is not an issue when you manually create inputs.conf.

Can I control how much performance data I index? I don’t want to go over the limit specified by my license?

Yes there are several controls to your JMX data. You can choose to collect performance, configuration, or both types of data. You can also specify how often you want to collect the data. For performance data you can choose the modules that you're interested in and choose which nodes those modules can come from.

• Was this topic useful?
• Post a comment