Monitoring enterprise IT has never been an easy task, and the speed, scale and complexity of digital environments is increasing exponentially each year. Digital complexity is not the future, it’s your reality—and it’s here now. IT departments are trying to visualize a moving target. The technical headaches alone can seem formidable: SD-WAN means that the unruly Internet is becoming your network, the cloud is becoming your data center and an ever-growing brood of SaaS applications have become your application stack. Your monitoring portfolio is heaving under the tensions caused by these complex changes. The natural instinct is probably to dig for a technological silver bullet, but the harsh fact is that legacy technologies and processes are not designed for these digital challenges. Time to take a good, hard look at the facts and detail some actionable solutions!
Here are the monitoring myths that need debunking (and fast!) to ensure your business has a digital monitoring strategy that is fit for purpose in the digital age:
Myth #1: The Internet is a “Black Box” that I can’t understand. I have to accept that I can’t do anything about outages.
The Internet is a “best effort” concern, which relies on many external providers—ISPs, CDNs, DNS providers—all of whom are constantly adjusting routing paths. For example, a single BGP routing decision can send traffic piling into the Great Firewall of China, bringing down multiple online services across the globe. Cloud adoption is extending your infrastructure risks beyond your firewall. Yes, the Internet is chaotic, and yet it’s core to your business.
But all is not lost. While you might not be able to find and fix an Internet issue directly—because you don’t own it—you can “evidence and escalate” the situation to someone who can. When a CDN, ISP or DNS vendor is shown concrete evidence that they are having an impact on an outage, they tend to fix it quickly. Actionability of insight is key here.
Recommendation:
- Keep a vigilant eye on Internet outages in real time: This is why we built Internet Insights™, a service that detects and helps solve service provider network outages using the power of collective intelligence.
- Ditch your old “find and fix” instincts, adopt an “evidence and escalate” mindset to keep the providers in your online ecosystem honest regarding their performance. Read our previous DNS report for data points that can help you choose the optimal Managed DNS provider for your business.
- Ensure your own house is in good order. Brush up on your best practices for CDN and DNS.
Myth #2: The native monitoring capabilities provided by my cloud mega vendors are “good enough” for strategic cloud monitoring.
Relying only on these tools for monitoring the full extent of your cloud environment can result in significant blind spots. The cloud mega vendors all offer the benefit of visibility into the internal workings of their respective platforms—CPU cycles, network packets, etc—but while their data can be useful, it’s only relevant within the environment the vendor controls. What they don’t give you is insight into the Internet plumbing, such as BGP, that their cloud service depends on. This becomes even more critical when your cloud traffic leaves their backbone, moving beyond their control (as it often will).
For example, in 2018, AWS connectivity was disrupted by a BGP hijack, which impacted Instagram services for several hours. And here’s a global example: Earlier this year, China Telecom’s network connectivity issues created a five-hour outage, which had a ripple effect across Internet connectivity, causing service disruption to Microsoft, Amazon, Workday, SAP and over 100 online services. You have to be ready to monitor outside of your cloud vendor’s environment to see and react to all of your online dependencies. The cloud-native tools provided “out of the box” only help you with a sliver of your total Internet and cloud real estate.
Recommendation:
- Don’t just rely on the native cloud monitoring tool. Ensure you have visibility into DNS, CDN, BGP, etc. that is transporting your cloud data. Don’t just trust, test!
- Be sure to test the performance of your cloud provider in specific regions, as needed. As our latest Cloud Performance Benchmark report found, performance can vary immensely between providers, especially across different regions.
- Educate yourself on the different connectivity and performance architectures that are being deployed by the major cloud vendors. Their backbone strategy will impact the quality of your digital delivery.
Myth #3: SD-WAN technologies provide greater visibility than my old WAN tech ever did. I don’t need extra monitoring capabilities!
When they start transitioning to an SD-WAN environment, many IT teams are understandably excited by the flexibility that SD-WAN solutions provide over a traditional WAN setup. And yes, it’s certainly a big step up! But SD-WAN solutions are not meant to be monitoring solutions. While some of the dashboards that SD-WAN management interfaces provide are nice, they don't give you full end-to-end visibility or a complete picture of how the Internet might be impacting SaaS and data center application performance.
TL;DR—By its very nature, SD-WAN typically means using less MPLS, and more Internet across your network means that you’re exposing enterprise networking to a highly unpredictable environment. And it can get even more complicated when you throw in Direct Internet Access (DIA) from branch locations, secure web gateways and other services into the mix. Even when opting for a more MPLS-heavy approach, the introduction of software-driven changes within your infrastructure will increase complexity.
Key questions to ask your team when planning SD-WAN:
- What are the performance characteristics of the ISPs you're planning to use, and will they meet your needs when you flip the SD-WAN switch?
- Do you know which ISPs are best suited to support particular SaaS solutions from different branch locations? Remember that not all providers deliver the same level of performance. Regional performance can vary markedly.
- What are the key network and application performance KPIs that will show rolling out SD-WAN has been a success? How do you plan to perform those measurements?
SD-WAN is one of the most consequential network architecture changes you will make—pushing cloud-driven changes deep into your enterprise infrastructure. In the cloud, there's no such thing as a steady state. If anything, SD-WAN means you need *more* monitoring, not less!
Recommendation:
SD-WAN success relies on not just having the right SD-WAN vendor in place, but also ensuring you have a bulletproof digital ecosystem (ISPs, CDNs, DNS, etc.) lined up to support the long tail of your deployment. Whichever SD-WAN approach—and vendor—that you decide to settle on, you will need to gather performance data before you roll out to production. This can help you to:
- Run smart PoCs when picking SD-WAN solution and services vendors.
- Focus on establishing operational visibility and firm KPIs in place before roll out.
- Have a data-driven process in place to reduce performance risk.
Learn about SD-WAN best practices in more detail on our blog.
Myth #4: I need thirty solutions to monitor the totality of my IT environment.
Over the past decade, monitoring stacks have become increasingly bloated. We’ve heard of instances where thirty to forty tools have been deployed to monitor IT environments. Unsurprisingly, enterprise IT teams see this as unsustainable and are running much-needed audits to consolidate down their portfolios. We’re increasingly seeing enterprises collapse their monitoring stacks down from multiple tools to around 5-6 strategic solutions. Often, this will be a combination of:
- AIOps
- Network performance monitoring
- Application performance monitoring
- Digital experience monitoring
We’re frequently chosen for digital experience monitoring due to the unique capabilities we offer for Internet and cloud monitoring.
Recommendation:
- Educate your teams on the requirements for a cloud monitoring stack. This will offer wider insights into cloud ecosystems and their dependencies.
- When consolidating your monitoring stack, be sure to keep critical SaaS apps like O365 and Salesforce top of mind. They behave very differently to traditional on-prem apps and RUM code injection won’t work in a SaaS environment. Ensure your chosen vendor can effectively monitor their architectures.
- Learn from a peer. Schneider Electric’s journey migrating to Salesforce Lightning is well worth hearing.
While the above monitoring “myths” are common, they are just the tip of the iceberg, and there are many others we could talk about.