3 Resolutions for Monitoring in the Cloud
Many network administrators still manage networks within their local data centers, but they’re increasingly troubleshooting in the cloud as businesses deploy more cloud workloads.
According to RightScale’s latest “State of the Cloud” survey, 85 percent of enterprises have a multi-cloud strategy. Businesses run an average of 41 percent of workloads in public clouds and 38 percent in private clouds.
The challenge of maintaining visibility over a complex network, in which virtual devices are provisioned and de-provisioned all the time, has made network monitoring tools essential. When your users experience performance problems with a cloud application, the problem could be within your cloud environment, in your internal network, or on the open internet. Without cloud-based network monitoring, problems are tough to pinpoint, especially when you have limited visibility into an outside cloud service provider’s network.
No one has to tell you that cloud is changing network administration and that it’s pushing you to evolve new skills. These principles can help you improve cloud troubleshooting and stay ahead of the changes.
Additionally, these three resolutions can help you to effectively retain visibility even in complex cloud environments.
1. Go Beyond Alerts
In the cloud, by the time you receive a network alert, the virtual device that generated it could already be decommissioned. It’s critical to log a lot of information about the application and device performance and to ensure that those logs remain, even when a cloud resource no longer exists. You also need logs that are generated before an error or warning happens. Logging helps you track down the conditions that created your network problems.
Extensive logging through network monitoring also provides your development team with a baseline of network and application performance. This baseline helps them automate infrastructure that is less likely to break. Sure, they can automate cloud resources for their applications without you, but the moment something breaks, they’re still calling you. If you’re alert dependent, you won’t have the forensics you need to fix their applications.
To take it a step further, adding active or synthetic testing to your arsenal is another way to not only react when an application goes awry, but occasionally to be the first one to detect it – not your users.
2. Become a Policy Wonk
Many network problems originate when cloud application-driven policies cause problems with network operations. Application policies have to play well with one another while also not derailing your security policies. Automation and orchestration tools may execute the policies developers put in place, and they may provide historical data on the resources applications need (server, storage arrays, etc.), but when the SLA isn’t being met, they need you to troubleshoot the policies based on an in-depth understanding of how devices work and interact.
As a 21st-century network administrator, you need to have some software-centric tools in your toolkit. It’s time to gain functional knowledge of the code that makes SDN and DevOps happen. You need to know Linux. You need knowledge of Puppet, Chef and other orchestration tools. You also need to become familiar with Docker and OpenStack. You don’t have to become a developer yourself, but you have to adapt to an SDN world. Deliver value by pointing out coding bugs and problematic policies that disrupt network operations, not only in the aftermath of an incident but also (even better) before they happen.
3. Monitor and Adjust
Today’s cloud environments, including the network, are too complex to troubleshoot reactively. Read the 2017 Gartner Network Performance Monitoring and Diagnostics (NPMD) Magic Quadrant Report and find out how why Gartner recognizes VIAVI as a leader in cloud-based network monitoring.
BONUS: Take Your Skills to the Next Level
Storm the castle, tune performance, or stay in your own orbit. Your skills and strategies reveal your IT alter ego, unlocking an all-new network performance monitoring power for 2018.