Welcome to Virtualization and Beyond
The Firm Foundation for Monitoring: Why it's all about that Baseline
Written by Jared Hensle, Senior Product Marketing Manager, SolarWinds
It's all about that base, that base, that baseline. Spectre and
Meltdown are bad, and when (if?) the final fixes are available to the general
public, you can bet your servers/applications are not going to perform as well
as before. But that is still one of the many great unknowns with these security
flaws. After applying these hot fixes and patches, how well are our operating
systems and applications going to perform? For a hint at what some individuals
and companies have experienced after applying "fixes, " you can read a few of
the blog posts that are currently being published on the subject. Some have experienced
more degraded performance than others, but what they ALL had in common were a
baseline and benchmark PRIOR to apply and fix.
How are you going to know what your performance degree is if
you don't know what your current state is? Without proper baselining, you're
left guessing that your websites, email, and custom applications may be running
slower than before. You're also biased from all the reports/blogs (such as this
one), and assuming performance is worse than it actually may be. What if you
applied these updates just prior to month-end accounting activities on your
accounting/database server? Your performance metrics are going to be completely
skewed vs. a random Tuesday afternoon. Without proper baselining, you don't
know that your GL reports add another x % to CPU and DISK I/O.
[ READ: How to plan for disaster recovery ]
Baselining is the foundation of any good monitoring
solution. Once you have gathered enough metrics, you can determine "what's
normal." Knowing what metrics, be it applications, systems, code, etc., are and
what they are expected to be at that given point in time allows you to
determine what's normal for your applications.
Baselining is not a once and you're done activity. As your business
grows your applications and data will change, so you need to be vigilant to
continue to monitor performance and continue to baseline. Let's say you add two
new applications with 100 concurrent user connections to a server. Of course,
you want to have the performance data and baseline information prior to adding
the new applications, but once it's done you need to look at re-baselining your
system. The prior alerts and metrics are no longer valid, and you and your
monitoring solution need to adjust accordingly.
Baselining also helps in scaling your organization, which
allows you to report on application and infrastructure changes. Monitor and
report to see if applications, operating systems, hardware updates, and other
software and infrastructure updates had a positive or negative impact on
performance.
Regardless of the number of fixes/patches required to fix Spectre
and Meltdown, you need to baseline your environment to see the business and
performance impacts that happen over time.
Read more articles from the Virtualization
and Beyond Series.
##
About the
Author
Jared Hensle is the senior product marketing manager at SolarWinds. In
his role, he is responsible for overseeing product strategy and execution of
the SolarWinds Virtualization Management business. Throughout his more than 15
years of experience as an IT professional, Jared has worked for multiple large
enterprises and run MSP and IT consulting firms.