Configuring a monitoring connection

This OctoPerf’s tutorial number 5 guides you trough the following steps:

  1. Selecting the previously created on-premise monitoring Agent,
  2. Configuring the Apache Tomcat connection,
  3. Selecting the counters,
  4. Tuning the default thresholds.

Notes:

If you already use an Application Performance Management solution such as Dynatrace or AppDynamics, you may prefer to integrate these APMs in OctoPerf.

Transcript

0s

Hello and welcome to the fifth tutorial of the new OctoPerf version.

4s

This is Guillaume speaking and today we will setup the monitoring of one of our servers.

10s

We will go back to the project we created in the previous tutorial videos.

14s

But this time we are headed to the monitoring tab.

17s

This is where we can select the agent to use and the type of monitor we want to create.

23s

It is important to note that all our monitors rely on system components,

27s

they will not generate any overhead on the servers themselves while monitoring.

32s

For instance Tomcat monitor uses the built in JMX, linux relies on system command, etc...

39s

Let's create a linux monitor with the agent we installed last time.

43s

On the first screen we have to enter the server IP or name and our credentials.

49s

A quick check tells us it is okay.

52s

Then the next screen is about selecting the applications.

57s

Applications refer to many things, the idea here is to let us select which elements (disks, network or processes) we want to monitor.

66s

Because next we are going to apply a monitoring template to these elements we selected.

71s

This template comes with a preselection of counters, that way if we are not familiar with one of these technologies,

78s

we can still get relevant information.

81s

And also we built custom counters that process several information into one,

85s

like the used memory on linux that needs to take into account cached and buffered memory.

92s

On top of that we added thresholds at industry standard values so that we get alerted if anything is wrong during the test.

99s

All of this can be configured of course, by adding/removing counters, or on the next screen where we can add our own thresholds.

108s

For instance if we drag and drop a new threshold, we can define its severity,

113s

the values which corresponds and how much time the issue must last before we raise an alert.

119s

Once finished, we can just leave this screen to either go back to monitoring and define a new one or runtime to prepare the test.

127s

Note that on the monitoring screen we can interact with any existing monitor and duplicate it if we have a lot of similar monitors to define.

136s

Also there is a quick check in case we want to make sure all our monitors are working fine.

142s

And we can deactivate a monitor if we don't want it to run for the next tests.

147s

That's it for today, thanks for following, I hope this was usefull.

151s

Feel free to comment if you have any question and keep in mind the build in chat is the shortest way to contact us.

157s

Next time we will launch a test and look at the different metrics.

160s

Thanks for watching, take care, bye.