While Software-as-a-Service (SaaS) has been a mainstay of Enterprise IT teams for years, many businesses have struggled to ensure consistent performance for their SaaS applications. A critical reason for this is SaaS blindness (i.e. not understanding the underlying reason for SaaS outages and/or severe performance issues resulting in cloud applications degradation).
SaaS adoption continues to accelerate, “Gartner forecasts show that in 2019, 44% of cloud spend will be on SaaS creating visibility challenges for I&O teams that can impact customer experience, revenue and brand reputation.”1 So that awkward SaaS blindspot in your IT stack will impact end user experience—and it’s not getting any smaller, quite the opposite!
In this blog, I’ll explain how SaaS monitoring differs from legacy monitoring, outline key technical considerations for SaaS solutions monitoring, plus some early phase steps to consider when spinning up a cloud monitoring strategy.
Flying Blind in a SaaS World Can Be a Career Killer
Would you feel safe being a passenger on a plane that’s flying on a cloudy night without any GPS, radar or air traffic control support? Of course not, you wouldn’t want to take that risk. In an IT context, running a stable of SaaS apps without having any SaaS monitoring in place (while not risking lives) definitely places your career at risk. It’s the digital equivalent of flying blind. Sooner or later, your critical SaaS apps will crash and you won’t know why. No IT leader wants to have their CxO and CEO shouting at them because their business-critical O365 or Salesforce applications are down when a key company deadline is looming!
Remember that there are a whole host of potential obstacles ‘cluttering up the sky’ that your SaaS traffic has to navigate—Internet outages, changes to public Internet routing, regional Internet Service Provider issues, SaaS load balancing issues, firewall issues, and so on. With so many potential connectivity issues looming, SaaS blindness can be an existential career risk for IT leaders.
But don’t just take our word for it. In the report, “Use Monitoring for SaaS Despite Its Limitations,” Gartner predicts that “zero visibility will increasingly be viewed as unacceptable, and as a result, SaaS monitoring will become standard within virtually every medium-to-large organization’s IT operations toolset.”2
The writing’s on the wall: SaaS monitoring will become the rule, not the exception. If you don’t have one already, a SaaS monitoring strategy should be a high priority for your business in 2020.
Why Can’t My Legacy Tools Monitor SaaS?
SaaS monitoring solutions are fundamentally different from legacy APM, ITIM or NPMD tools. Traditional code injection techniques won’t work on SaaS services, because the SaaS vendor doesn’t permit it. SaaS monitoring is laser focused on determining the point of origin of an issue and who is impacted. This contrasts strongly with a legacy tool’s broader focus on finding "something to fix." This makes SaaS a "black box" as far as legacy monitoring is concerned.
As for what constitutes the core of a SaaS monitoring solution, in the report, “Use Monitoring for SaaS Despite Its Limitations,” Gartner suggests that the most critical features for SaaS monitoring technologies to include are:
- Cloud and on-premises synthetic transactions (with benchmarking)
- Internet performance monitoring (IP tracing, packet inspection, etc.)
- Depth and breadth of SaaS-specific sensors3
Our belief is that a SaaS monitoring solution should leverage holistic Internet outage tracking, which gives visibility into things such as BGP routing and ISP-driven changes. To refer back to our earlier airplane/flight control analogy, you need to be able to visualize the crowded (and ever-changing) “air space” that your SaaS traffic is flying through.
Diversity of Insight Matters: Avoid Only Monitoring from Cloud Vantage Points
For SaaS monitoring, it's crucial to have high-quality sensors in place, reflecting the reality of your users’ actual locations. The quality of insight provided by the sensor is also critical. Ideally, your sensors aren't deployed exclusively from inside a cloud mega vendor's infrastructure. Why not? By moving all monitoring vantage points into the public cloud (as some large monitoring vendors have chosen to do), the likelihood is that you deprecate sound visibility into digital experience delivery.
As we’ve pointed out previously, while the cloud mega vendors have excellent backbone infrastructure supporting their own cloud platforms, no end users sit in AWS, Microsoft Azure or Google Cloud data centers! You need to have visibility across tier 1, 2 and 3 ISP providers to reflect the totality of your users’ experiences. If your monitoring vendor decides to exclusively monitor from within public cloud data centers, that is a significant loss of visibility. You should also be wary of vendors who have inconsistent monitoring capabilities, depending on the type of agent.
Master Your Digital Reality: Embrace an “Evidence and Escalate” Model
Software has been the tip of the spear for digital transformation, and SaaS has transformed the shape of the enterprise. To put it bluntly, the cloud is becoming your new data center, the Internet is becoming your new network, and SaaS is becoming your new application stack.
From a workload perspective, your mission-critical business activity is increasingly likely to be sitting in a digital environment that you don’t own. In this digital environment, your ISP, SaaS, IaaS and cloud security providers will be pivotal in ensuring consistency of your business performance across your digital services. The digital ecosystem that your business runs on, sits outside your four walls. Crucially, it cannot be fixed directly by your team in the same way your legacy IT can be. As well as requiring new monitoring technologies, this new reality also requires a new monitoring psychology. In many cases, evidencing the origin of a problem and escalating to the provider best able to help solve it, may well be the only way to achieve resolution.
Getting visibility into digital SaaS-driven environments leads to a pivot from a reactive legacy “find and fix” mentality to a more proactive “evidence and escalate” mindset. Rather than sitting on a 100 person war room call trying to figure out why Office 365 is down, real-time SaaS monitoring can help you quickly evidence the origin of the issue to the relevant ISP, SaaS or cloud security provider who is best placed to help. And because you are approaching said provider with a data-driven view of the precise cause (rather than just raising a general accusation), they can take quick actionable steps—changing a routing policy, for example—to rectify response time issues. This is a fundamentally more proactive approach to ensuring a digital experience.
Kick Those SaaS Tires: Test, Test, Test!
Talking about proactivity: you can also use SaaS monitoring to kick the tires on a critical application prior to making a major update or undertaking a migration. For example, Schneider Electric, the global leader in energy management and automation, successfully migrated 45,000 employees and several hundred thousand customers and partners worldwide to Salesforce Lightning by leveraging synthetic monitoring and Internet visibility from ThousandEyes. Again, you can minimize risk and ensure your SaaS usage is smooth by planning ahead, with the right testing and visibility in place.
How Should I Think about Integrating SaaS Visibility into My Existing Monitoring Stack?
Our experience is that many enterprises don’t just look at SaaS monitoring in isolation, and they end up re-evaluating their monitoring investments more broadly. Typically, they decide to consolidate their monitoring stacks from as many as 30 monitoring tools to around 4-5 critical solutions. As well as SaaS, this process is driven by initiatives like SD-WAN and the shift towards digital platforms more generally. A consolidated “Modern Ops” stack tends to revolve around the following components:
- AIOps
- Network performance monitoring
- Application performance monitoring
- Digital experience monitoring
But the above list isn’t exhaustive, there’s no definitive approach to a digital monitoring portfolio. Your business is complex and the mix of technologies you will need for SaaS and digital monitoring really depends on what’s most important to your business success. As to where the SaaS component plugs in, some elements of SaaS monitoring can be found in NPMD and APM solutions, but here at ThousandEyes, we believe DEM is where the natural center of gravity sits for SaaS monitoring. Gartner’s monitoring taxonomy highlights that SaaS monitoring tools share a "core set of functions that largely fall into the taxonomy of DEM."4
Many customers specifically select us as their DEM solution of choice, due to the unique visibility we provide for Internet and SaaS environments. For more insight on Digital Experience Monitoring, read about it in this blog.
To wrap things up: Here are some opening questions you can ask yourself as you begin your SaaS monitoring journey:
- Is my network ready to handle Office 365?
- What’s the connectivity in key branch offices vs. remote workers?
- How many of my priority users are remote vs. onsite?
- What are the key application performance KPIs that will show rolling out a SaaS application has been a success? And how should I perform those measurements?
But these are just some of the considerations you should think about when it comes to SaaS monitoring, there are many more.
What’s Your Biggest SaaS Challenge for 2020?
If you need help solving a SaaS monitoring challenge, or would like to see the ThousandEyes platform in action, request a demo today.
References
1Gartner, Page 2 – Market Guide for Digital Experience Monitoring, September 2019
2Gartner, Page 11 - Market Guide for Digital Experience Monitoring, September 2019
3Gartner, Page 19 - Use monitoring for SaaS despite its limitations, September 2019
4 Gartner, Page 19 - Use monitoring for SaaS despite its limitations, September 2019