How cloud-monitoring services can help you avoid your own ‘Fail Whale’

We are excited to announce the first breakout panel at CloudBeat 2012, “Visibility: Cloud Management and Control in a Complex World,” featuring speakers from 6 Fusion, Cloudability and Appescute. 

Even when you buy a car with a warranty, you still have to get it serviced. Your cloud infrastructure is no different.

“It’s tempting to think of [the cloud] as something that’s already taken care of,” said Joshua Greenbaum, principal at Enterprise Applications Consulting, a Bay Area-based research firm. “Even if you’re a startup, you may still have to do a tremendous amount of sophisticated monitoring.”

Customers are notoriously fickle, and for digital natives, failure is not an option. From a small garage startup to a large enterprise, IT is under intense pressure to spot networking problems in seconds — before they disrupt customer service. That’s especially true for startups and small to medium-sized businesses (SMBs).

“For startups without much of an existing reputation, it’s even more critical to have reliable uptime for your website and services,” said Donnie Berkholz, an analyst at Redmonk, a developer-focused research firm. Berkholz stressed that the dangers for SMBs and young tech companies are very real, even if they’re not hosting a mission-critical application or providing a mission-critical service.

Cloud monitoring is a necessity, but the market is flooded with options. So VentureBeat asked a host of experts to cut to the chase: What are the best cloud-monitoring services out there?

Well, options for cloud monitoring tools will vary depending on the size of your company.

Small to medium sized businesses

Pingdom, a free web monitoring tool, is the most lightweight cloud monitoring solution on the market — it’s a simple ‘heads-up’ alert system that is popular with startups and SMBs. (Tip: Next time Twitter is down, you can track its performance on Pingdom.)

A simple monitoring service is not a cloud-management tool, but it’s a start. For more full-fledged options, you have a number of choices.

Rackspace recently announced cloud monitoring tools designed to let businesses monitor their IT infrastructures, regardless of whether they’re hosted in the cloud or on-premise. For Greenbaum, this is the most viable choice for startups and SMBs that are already using Rackspace for their cloud hosting.

“Our competitive advantage is how well it integrates with the rest of Rackspace,” said Pat Mathews, senior vice president of corporate development at Rackspace. “Beyond that it’s scalable and built for the new workloads.” Austin, Texas-based Rackspace offers a portfolio of cloud products and picked up Cloudkick in 2010, which forms the basis of the new monitoring tool. It works by sending customers an alert if things are going wrong with equipment, irrespective of where it’s hosted.

Of course, Rackspace isn’t the only choice.

Ray Wang, CEO at Constellation Research Group and an expert on all things cloud-related, says your decision depends not only on the size of your business, but also the vertical you’re in.

For e-commerce companies, Wang suggests AlertSite, an applications performance management solution with 100,000 users, acquired last year by Smartbear software.

“What you want is rich tools, easy analytics, and the ability to transcend cloud and on-premises,” he said.

Copper Egg, he told me, is another good solution for those that have opted for hybrid cloud environments. Copper Egg also won the approval of Berkholz.

Berkholz’s other top picks are Boundary, which offers speedier monitoring for hybrid cloud environments, and Zenoss, an open source option.

For larger enterprises

IT departments understand that legacy IT monitoring solutions are riddled with problems: They were never developed to deal with hybrid cloud environments. They are comprised of piecemeal technologies, originally designed to gain insight into whether services and applications are living up to their service-level agreements (SLAs) and flag IT when there are service interruptions.

Trying to configure a legacy solution to extend monitoring beyond the traditional data center presents a new set of cost, learning, and reporting challenges for CIO’s, as eWeek reports.

And when it comes to the big-boy vendors in the space, competition is intense. The experts throw up a smorgasbord of names, with BMC, Nimsoft by CA, and ScienceLogic popping up most often.

Amazon Web Services (AWS) is by far the largest provider of cloud-based infrastructure services. It competes with Microsoft Azure, Joyent, and Rackspace. Amazon has its own cloud monitoring tool called CloudWatch, which it offers to premium customers. If you’re a big user of AWS, this is one option to investigate.

This is a decision best left in the hands of IT. However, here is a series of questions big companies can ask before making a decision about a cloud monitoring provider:

1. Is the architecture scalable? Can it deliver monitoring that meets the expectations of the organization’s growth?
2. Does it deliver on-demand monitoring? Can it respond in a matter of seconds?
3. Does it meet your budgetary requirements? Are there flexible purchasing options?
4. Are there customizable dashboards that grant you visibility?
5. How speedy is the deployment?

It’s still not yet clear whether large-scale enterprises will stop dragging their feet and make the jump to a full cloud-based infrastructure. One way or another, they should be on the look-out for a monitoring tool that can keep up with the pace. It’s all about agility, flexibility, and speed.

As Gary Read, CEO of Boundary and formerly an executive at Nimsoft, put it, “Most IT management tools were designed in a different era when infrastructure was static. It’s no longer acceptable to take a look at your applications every five minutes.” That would be a massive fail whale.

Where to learn more

If you’re stuck, Christian Dawson, chairman of the board of directors at the Internet Infrastructure Coalition, recommends sites like WebHostingTalk.com, where you can ask peers and experts for their thoughts about various companies’ capabilities. “Nobody is going to have perfect feedback — the market is still too young, fragmented, and misunderstood, but you’ll be able to tell pretty quickly who is reliable and responsive,” he said.

If you’re new to all this, before you go shopping, the first decision you’ll need to make is a structural one. Will you build your monitoring in-house using open source software, or will you outsource to an existing company? The majority of the experts recommend the latter.

Finally, ask the sales guy the tough questions around things like actionable alerts. In a recent interview with Rackspace to discuss the launch of its cloud monitoring tool set, Pat Matthews, senior vice president of corporate development, admitted that this is a notoriously tricky problem for vendors. Some problems need to be rectified by a developer, others will fix themselves. So which alerts are truly actionable? “We are starting down that path,” he said. “It’s something we will continually strive to improve.”

The good news is that there isn’t a world of difference between the various solutions on the market. Of course, there are the legacy, best of breed solutions that EMC, HP and the like offer. But if you’re not running a large enterprise, you have plenty of cost-effective alternatives.

“With rare exceptions, we’re talking about very similar kinds of products,” said Greenbaum. He used the analogy of a heart monitoring system. “You don’t care about the brand or the packaging,” he said. “You just need something that works.”

Security camera image via Shutterstock