Table of contents
| Monitoring an Application Server or Database |
| Using Additional Load Generators |
| Scheduling a Test |
This tutorial describes how to create, configure and use a monitor for monitoring your application or database server. It sets out the steps involved using an actual example, the Linux monitor.
To gain the most from this tutorial, it is recommended to have read the reference documentation on "Monitors".
NeoLoad provides numerous statistics on your application, focused mainly on your pages' response times. It's often useful and instructive to combine and check those figures against statistics coming from components in your application architecture such as the application server or database server. NeoLoad provides features that let you monitor these components . This tutorial describes the different types of monitors provided in NeoLoad and where and how they can be defined, configured and used.
To start defining monitors, we'll first click on the
"Monitors" tab in the "Design"
section of NeoLoad.
To add a monitor, we first need to specify which machine we want to
monitor. To do this, click on the "New Monitored
Machine" button.
NeoLoad displays a dialog box that allows you to define the monitors to be placed on the machine:
The dialog box lists the types of components and/or protocols that NeoLoad is able to monitor. These monitors are optional modules; if you are interested in a particular monitor type, please contact Neotys for more information.
Enter the machine's host name or the server's IP address and select
the monitors you wish to place on the machine. For our example, we'll
select the Linux monitor.
Click "Next >" to configure the monitors you
have selected. The following panel displays the monitor
information.
Click "Next >" once more to enter the
monitor's details. For more information on the specific settings for
each monitor, see the Reference Guide.
To make sure the monitor is working properly and that NeoLoad can
connect to your server, you may test the connection by clicking on the
"Test" button. If the monitor is working, the message
"The connection settings are correct" is displayed.
If not, details of the error are displayed.
Click "Next >" to define the performance
counters you wish to put in place on your server. Depending on the type
of monitor selected (in our case, Linux), NeoLoad
will automatically define the most appropriate counters to be monitored
and the associated alert thresholds. Check or clear a box to add or
remove a counter from the monitor.
Here, we have added the CPU User, CPU
System and CPU Idle counters, among others,
to the Linux SSH monitor.
Click on the "Alert thresholds" tab to set the
alert thresholds for the desired performance counters.
In this example, two thresholds have been set for the
"CPU Idle" counter:
a warning-level alert will be triggered if the counter's value falls to or below 50 for more than 5 seconds.
a critical-level alert will be triggered if the counter's value falls to or below 10 for more than 5 seconds.
Then click "Finish" to create the monitored
machine, monitor and associated performance counters.
While the scenario is running, you may select the
"Runtime Monitors" tab in the
"Runtime" section and add any monitor information you
wish to the display area.
The "Results" section allows you to view the
graphs representing the various items monitored during Runtime. The
screenshot below shows a graph obtained by combining user load and the
Linux SSH monitor's CPU Idle
counter.
In this example, the value of the CPU Idle
counter regularly exceeds the alert threshold <=
50% at a user load greater than 120 simultaneous virtual
users.