Another Day. Another Product.

October 18, 2010

Today, SoftLayer released an Advanced Monitoring Solution based on Nimsoft's Monitoring software suite. In a nutshell the product will give SoftLayer visibility onto a customer's server at the OS level. In addition to the great product benefit the customers receive, it will add tools to our sales and support staff to troubleshoot, diagnose and systems design.

The core product works through a piece of software that gets installed on a customer server called a robot. The robot in turn allows probes to be run on the server. The different probes collect various data points from the OS and applications. As the probes collect data they pass the information onto some intermediate backend servers, and eventually end up on our brand new HBase data warehouse (HBase is the massively scaled database for large amounts (petabytes) of data). This is the corner piece for the scalability of the offering. So, robot is the main software and the probes are the application watchers that run inside the robot framework.

There are additional features outside of the process mentioned called 'Offbox Monitors' or 'Offbox Probes.' These are probes that live on servers in the SoftLayer data center. The idea behind these is that we are able to let customers have network services they want to monitor from a remote location. An example would be url_reponse, which monitors if a url is active and passing data (along with some other pieces of data people might be interested in like response time).

What it can monitor? The better question might be what can't it monitor? The SoftLayer offer comes in three packages - Basic, Advanced and Premium. Basic is a free package that monitors core hardware (CPU, memory, disk) along with simple process and services. Advanced moves into deeper system monitoring, and Premium adds more application monitoring (like databases, web services etc.). This offering is available for hardware, Monthly CCI's and hourly CCI's - basically for everything we sell. Customers can order the software from all order forms (external, internal, cci, server etc.) as well as add the service post deployment from the advanced monitoring portal page.

The service offering has two distinctive reporting features that we call graphing and alarms. Graphing allows customers to (yep, you guessed it) graph all the data we collect. For example, we can show a graph of CPU usage over time. Alarms are notifications that services are outside of a predetermined range. For example, you could setup an alarm when CPU usage goes above 90%. Alarms can be tracked from the customer portal, or email alerts (SoftLayer calls this list 'Alarm Subscribers') can be setup by the customer.

All the features of the product are accessed via the customer portal, or via SoftLayer's internal portal. Configuration, graphing, and alarm management can all be done from one management page in the customer portal. In near real-time customers can change configurations directly on their server or cloud computing instance (CCI) for the various data points they want graph and alarm. It's pretty slick, and it adds to the SoftLayer secret sauce. We have also added a feature that allows the customer to save configurations on a particular server and “redeploy” them on different servers or future servers they may add. This feature makes it easier to scale and customize for a particular customer's needs.

As time goes on we will continue to add more probes and more features. This is just the beginning - make no mistake it's pretty damn cool.

-@quigleymar