If you’ve run across other articles I’ve written about hybrid cloud, you’ll notice a familiar theme: it’s an artifact of other technology decisions nobody asked for. As such, it’s left up to IT operations to figure out how to know if it’s working well, cost-effectively, or as expected in the first place. However, tools are improving rapidly, and IT teams are gaining skill with the model and technologies. The question then is where are most enterprises on the spectrum, from realizing they’re in a newly challenging, hybrid on-prem, cloud, and SaaS environment, to getting it under control? Recently, I had a chance to take this conversation to Twitter.
Theresa Miller, Instigator
Theresa Miller (@24x7ITConnect) is a longtime application ops expert as well as Citrix CTP, VMware vExpert, and Microsoft MVP. She knows a thing or three about apps. Moreover, she’s helped companies transition through several epochs of application architectures, with cloud the most recent. She wrote about some of the particular
challenges of monitoring in hybrid cloud environments, with a focus on how monitoring is evolving to take some of the headache out of hybrid cloud ops.
The comments were interesting, with readers reporting they have increasing confidence managing hybrid cloud. That’s reassuring. Unfortunately, some also shared the rate of hybrid and cloud-native tech complexity is increasing at the same rate, only partially clearing skills gaps and tools debt. To open up the discussion, we took a few minutes to have
a chat about it on Twitter and follow-up Theresa’s article in more detail.
280-Character Takeaways
Looking over the bite-sized chunks of recommended practice from the conversation, I’ve rolled up four main takeaways. Although each hybrid cloud ops team is as different as the combination of vendors and tech they manage, the basic challenges of hybrid apply to nearly everyone.
- Hybrid has multiple sides. Organizations may have ended up running a hybrid cloud for many reasons. It could have been a specific business requirement, a cost-saving measure, or even a “modern team in the cloud” who suddenly finds themselves with a need to follow transitions back to on-prem or wherever back ends are deployed.
- Vendors made hybrid sound easier than it is. Digital transformation, modernization, lift-and-shift sounded straightforward-ish enough, but the reality of maintaining environments with more complexity and components that may break, is less so. If your team was surprised by new requirements for investment, you’re not alone.
- Troubleshooting hybrid can be substantially different. Because of the nature of new technologies introduced in hybrid cloud adoption, tried-and-true monitoring methods may not work. It can take a while for teams to regain the level of multi-system visibility and control they had when everything was in an application stack in the data center. Step one is to get unintended monitoring system(s) complexity under control and into an integrated view. Nobody likes monitoring multiple monitoring systems.
- Unanticipated costs drive interest in hybrid cloud monitoring. By now we’ve all realized cloud isn’t cheaper by default and are still looking for the sales reps who claimed early on it would be. Downtime remains the key business cost, but for IT OpEx, the specific expensive details of each environment can be very different. Worse, they’ll be recurring until the team can identify overprovisioned, unused, misapplied, and unsuitable resources in an ongoing, automated fashion. Intentionally or unintentionally, cloud is designed to let unnecessary charges accumulate over time.
Déjà Vu All Over Again
Hybrid cloud challenges are neither the first nor the last examples of the unintended effects of technology transformation. Major change always starts with developers and vendors, eager to find new or better ways to deliver the services users expect. As such, tools, especially mature and accessible ones are usually the last components delivered—features first,
then management. Teething pains are a constant irritant to IT, but we generally adapt quickly.
In this case, however, there are often more new elements to monitor and manage at once, and sometimes completely outside the experience of all but a handful of experts. You might think MongoDB is similar enough for the Oracle DBA team to take it over, but that’s not true. And Mongo is only one of many typical new elements of the new hybrid app including Redis/NGINX/Kubernetes. Each has its own learning curve, new error modes, and specific resource requirements. But once again, IT teams are showing good progress getting hybrid cloud under control. That shouldn’t be a surprise to anyone who knows technology teams—it’s what we do every day.
