
Welcome to
Virtualization and Beyond
Contributed by Michael Thompson, Principal, Product Market Management, SolarWinds
Monitoring in Hybrid Environments
The cloud needs no introduction. It's perhaps one of if not the
most pervasive topic in IT today, and the pace of adoption of cloud-based
technologies is rapidly increasing. Some companies have even moved to a "cloud
first" strategy. However, for most companies the bulk of IT activity still
involves monitoring and managing on-premise systems. Even still, there is a
marked shift across the board towards the cloud. For those companies with the
bulk of IT still on-premise, that means a hybrid on-premise and cloud
environment to at least some degree.
As this shift occurs, though, the monitoring and management
of these hybrid environments is often left as an afterthought. This is a
mistake that can have long-lasting repercussions. To avoid
this, here are some considerations to take into account when thinking about
monitoring hybrid environments.
Monitoring Objectives Shift
With both
cloud and hybrid environments, the foundational objective of the monitoring and
management systems quite often changes. The traditional focus of on-premise IT monitoring
and management has been first and foremost to ensure that individual components
and systems - servers, applications, networks, storage, etc. - are available
and performing, and if they aren't, to quickly pinpoint the problem and fix it.
That can
fundamentally change with the cloud portion of a hybrid system in that instead
of pinpointing
where and
what the problem is, the primary
objective is to be able to quickly and definitively determine
who owns the problem. This can require a
different way of thinking about monitoring that requires clearly delineated boundaries
and data requirements.
Start With the Business Objectives
With
distributed ownership of the resources delivering business services, it is more
important than ever that the design of a management system starts from
understanding and monitoring the end business service or objective that the
hybrid system is delivering. Whether the hybrid system drives sales leads,
e-commerce revenue or internal production data, starting with the ability to
identify problems with the end service before users start complaining is key.
From there,
the next step is to work backwards to understand the components required to
deliver that business service. If a range of services and infrastructure from
both on-premise IT and off-premise cloud providers are required, you must be
able to quickly isolate which one is not performing appropriately. Essentially,
the idea is to trace back a new troubleshooting plan before a problem arises to
ensure that when a problem does present itself, you have the data you need to
determine where the problem exists.
Different Ownership Requires Different Data
If the cloud
service is essentially SaaS, where you have no management of the technology,
it's pretty much a matter of understanding what failure would look like and
ensuring there is a SLA and that you have appropriate metrics to show the SaaS
provider for a fix if needed. However, in situations where the cloud provider
is delivering infrastructure and you are installing and running software on top
of it, it's a little trickier.
The first thing
needed will be to determine if it's your software or configuration that is
causing the problem or if the cloud provider is not meeting the SLA or otherwise
delivering inadequate performance. To achieve this, a monitoring system must be
in place to definitively determine where the problem lies. The typical, "I
don't see any problems in my environment, must be on your end," will likely not
be enough to force a cloud provider to take action.
Remediation May be Different
When it
comes to on-premise infrastructure, responsibility to fix it when it breaks
lies solely on IT. One of the major advantages of the cloud is that if a
problem with a VM or other cloud component arises, it can simply be killed and
re-provisioned instead of fixing it. However, this requires applications to be
architected appropriately so the business service provided by the application
isn't impacted.
Make Friends with Development
Many of the
monitoring decisions need to be settled up front, when the hybrid or cloud environment
is being created. In an on-premise only scenario, developers have little choice
but to involve operations teams because they have to coordinate with them
regarding the hardware that their applications would end up running on. With
cloud resources, though, developers with a credit card or purchase order can
get the infrastructure themselves. Without coordination, operations can end up
with the responsibility over applications that aren't designed for cloud
remediation, without monitoring that determines who a problem belongs to and with SLAs that aren't useful in
maintaining business objectives. The easiest way to prevent this is to
proactively reach out and engage with development teams so all are included in
the early phases of development.
Plan for the Worst so It Won't Happen
Thinking
through and planning for some of the worst case scenarios in hybrid
environments early on can prevent problems before they come up. By setting up appropriate
monitoring systems, designing applications for remediation and obtaining
applicable SLAs with cloud providers, operations teams can ensure business outcomes
for hybrid environments just as we've done for years for on-premise environments.
##
About the Author
Michael Thompson, Principal, Product Market Management, SolarWinds. Michael has worked in the IT management
industry for more than 13 years, including leading product management teams and
portfolios in the storage and virtualization/cloud spaces for IBM. He holds a
master of business administration and a bachelor's degree in chemical
engineering.