How to get better visibility into cloud app performance
- By Joel Dolisy
- Sep 29, 2015
With virtualization now firmly entrenched in the data center, many agencies are struggling with the next big thing — migrating applications to the cloud and monitoring performance and availability.
After all, applications play a critical role in helping agencies deliver services and information to the public, employees and warfighters, and any negative impact on performance can have far-reaching effects. A recent SolarWinds survey found that 93 percent of end users reported that application performance and availability affects their ability to do their jobs, with 62 percent saying it is absolutely critical. In addition, one in five said slow or unavailable applications result in significant financial impact for their organizations.
While the cloud does offer cost, flexibility and agility benefits, issues with performance, requirements for mission critical applications, privacy, compliance and security and the sunk cost of existing application purchases are valid concerns. Thus, rather than going all in on the cloud, most agencies will likely manage a mix of cloud-based, on-premises applications and other IT infrastructure elements -- or hybrid cloud -- for some time to come.
While a hybrid cloud approach seemingly offers the best of both worlds, it brings with it a new set of complexities and challenges when it comes to ensuring application performance, which in turn require particular best practices to overcome them.
A transition to hybrid cloud environments means IT is now responsible for two sets of infrastructure: the cloud deployment as well as traditional, on-premises systems, both of which must be managed to ensure applications operate at peak performance.
Further compounding the challenge is that administrators typically don’t have the same level of visibility — or any, in some cases — into the performance of cloud environments. If there is a performance problem, purchasing more storage could help, for example, but there’s usually little to no visibility into what’s happening to the cloud infrastructure to know if that’s the best solution. And while it is possible to move to higher-performing cloud infrastructure, that can be expensive -- and upgrading an entire environment is often overkill.
Seeing through the fog
Even traditional, on-premises applications can be difficult to manage. According to the survey mentioned above, 81 percent of respondents contacted their IT department in the past year over application performance or availability, with one-third having done so six times or more. Furthermore, more than one-third said they waited over a day for problems to be resolved, with 22 percent waiting several business days for fixes to critical applications.
The reason application performance management is still so hard is because applications are becoming increasingly complex, relying on multiple third-party components and services that are all added to the already-complicated modern application delivery chain that includes the applications and the backend IT infrastructure that supports them —plus all the software, middleware and extended infrastructure required for performance. This modern application stack, or AppStack, is often managed the old-fashioned way — in silos, without holistic visibility of the entire stack at once.
When the cloud gets added to the mix, none of the traditional AppStack goes away, but it’s no longer managed by the hosting agency in an on-premises datacenter. Nevertheless, IT is still accountable for application performance. So, what can application administrator do to ensure consistent top performance in a hybrid cloud environment? Here are a few things to consider:
Add a step. In a traditional on-premises IT environment, the first step in troubleshooting an application performance issue is to ensure that individual components and systems of the AppStack -- servers, applications, networks, storage -- are all available and performing properly. In a hybrid cloud system, however, instead of pinpointing where and what the problem is, the first step is quickly and definitively determining who owns the problem -- IT or the cloud vendor.
Get involved in the development phase. Cloud monitoring and management decisions need to be made when the hybrid cloud environment is being created. Because a cloud environment allows developers to access the infrastructure for their applications without consulting the IT operations team, IT pros are often left to manage applications that aren’t designed for cloud remediation and do not have the monitoring capabilities necessary to determine who owns an application performance problem (see above). By getting involved in the development phase of the hybrid cloud environment, administrators can ensure they have control over application performance.
Manage cloud and on-premises performance data to determine the root of the application issue. Application performance management in a software-as-a-service (SaaS) environment requires IT personnel to understand what the application failure looks like, and to ensure there is a service-level agreement (SLA) as well as the appropriate performance metrics to show the SaaS provider that there is an issue. In an infrastructure-as-a-service (IaaS) environment, managing application performance gets a little trickier. IT needs to determine whether the application performance issue is with the software or the configuration. To achieve full visibility into the performance issue, a monitoring system must be in place to definitively determine where the problem lies.
Plan for the worst-case scenario to prevent it from happening. Early on, administrators should think through and plan for the worst-case scenarios in a hybrid cloud environment to spot problems before they arise and to be prepared should they actually occur. Make sure, for example, that critical systems can failover to standby systems and data is backed-up and archived according to agency policies.
Every technological shift comes with a unique set of complexities and challenges – and the need for new understanding. The hybrid cloud is no different. However, gaining visibility into application performance can help agencies reap the benefits of hybrid cloud environments and ensure application performance remains consistently strong.
Joel Dolisy is the CIO at SolarWinds.