Are you really getting the cloud uptime you expect?

One of the scariest things about putting information in the cloud is not knowing if you’ll always be able to access it. Companies want outstanding uptime, and providers promise it. But who’s checking if they’re actually delivering? 

82557029

A recent study by CloudEndure shows that there’s a bit of a disconnect between what companies think about their cloud providers and what they actually know for certain. When asked how much uptime they were aiming for, 79% of respondents said they wanted “three-nines” or better. That’s 99.9% uptime, or less than nine hours of downtime a year.

They thought they were hitting their goals pretty well. Respondents said they met their goals:

  • most of the time (49%)
  • consistently (43%), and
  • some of the time (7%)

Only 1% of respondents said they were unable to meet their goals.

Not all tested

Sounds pretty good, right? The only issue: Many of those respondents don’t even know if they’re actually getting their money’s worth. When asked about their methods of calculating uptime:

  • 40% said they use their own metrics
  • 34% use third-party tools or services, but
  • 26% don’t measure at all.

So more than one-in-four respondents doesn’t calculate uptime in any way. So while they claim they’re hitting their goals, many just flat-out don’t know if that’s the truth.

This isn’t some “gotcha” question. A full 29% of the IT pros surveyed rated availability of services as “most critical.” Nearly 7 in 10 (68%) ranked it as an 8 or higher on a scale of 1-10 in terms of importance.

And outages aren’t a hypothetical situation for most of these respondents, either. More than half (54%) said they’d experienced an outage in the past three months. A quarter (28%) had an outage in the last month.

Only 17% were lucky enough to have never had an outage.

It’s not easy

Clearly, if there was a simple solution that let everyone maintain uptime of 99.9% or better, it would be industry standard. No one wants to have unscheduled downtime.

So what’s holding companies back from improving on availability goals? Some familiar roadblocks made the top 3:

  1. insufficient IT resources
  2. budget issues, and
  3. difficulty preventing software bugs.

The most likely culprit of these outages were application bugs, network failures and human error.

Double-check promises

It’s said you can’t improve what you can’t measure. And there’s no arguing that point.

Make sure to:

  • Track the uptime. Every time there’s an outage, make sure to add up that time and keep detailed records of how long it took to get back up and running. Have clearly defined metrics for what “downtime” is, and what a recovery looks like. It may be partial access to the cloud or full access depending on your situation.
  • Compare with providers. See what your providers have promised or said that they’ve delivered. If you’re getting wildly different figures for uptime, that’s a red flag. Also, make sure to see how they define the same terms that you do to avoid miscommunication.
  • Check contracts. Find out if your contract allows for any recourse if they don’t deliver on uptime promises (or if it even addresses the issue at all). This may be something you want to build into future contracts from here on out.