It’s been a few weeks since I got back from trekking to the Everest Base Camp in Nepal. Reaching an altitude of over 5500 meters (18,208 ft), and seeing the spectacular gold-bathed peaks of Everest, Nuptse, and the grand Lhotse at sunset made the grueling 10 days leading up to the views well worth it.
But apart from the spectacular views, the two weeks in the mountains taught me a “higher” appreciation for the things we often take for granted: hot showers, warm clothing, and fast and reliable internet and cellular networks. Dealing with weak cellular data connectivity in the highest mountain range on earth has taught me to further appreciate faster loading webpages and the impact they have on human lives. It made the importance of website performance monitoring even more apparent to me.
Do I sound dramatic? Probably a little. The trekkers on the Himalayas using a 2G or a 3G network may make up an extremely small percentage of a website’s end users. But, if the webpages load fast and are responsive for this demographic, then the performance engineering teams are doing their job.
As a performance engineer, I’m always drawing parallels between my daily actions and web performance. Here are my takeaways as it relates to website performance monitoring from my fortnight on the mountains:
- Availability trumps performance, barely
- Data is expensive so optimize for mobile networks
- Anticipate failures and get a good insurance plan
- Reaching the summit is not the end, continuous monitoring is necessary
- The marriage between IT and Business must be a stable one
I expound on these observations below:
Availability trumps performance, barely
So often the web performance monitoring community talks about how a slow loading website is still better than an unavailable website. At 15,000 ft., this could not have been truer. When access to the internet or network is sometimes available only for a few hours in the day (with solar-powered cellular towers for instance), I would have taken a slow loading webpage anytime. But then once I got on the network and faced a webpage taking eons to load, I would have rather just put down my phone and hunkered down next to one of the heaters.
So yes, it’s essential that your webpage is always available. Downtime equals revenue drop and a negative impact on the brand image. And this makes high frequency availability monitoring an indispensable part of any website performance monitoring strategy.
But studies have also indicated an even higher revenue drop or a higher chance of shifting to a competitor for consistently slower websites when compared to an occasionally non-available webpage. Hence, monitoring your website for performance degradations becomes almost equally as important.
Data is expensive so optimize for mobile networks
With the influx of unlimited data plans, we rarely tend to appreciate the luxury of all that data. But when you’re forced to pay $10 for every 250MB of data, you become extremely aware of every byte’s consumption.
For example, imagine trying to post about the biting cold (-25 degree Celsius, -13 Fahrenheit) on all your social media apps. This would definitely set you back by a few MB and that would be enough to make you want to reconsider— is it cold enough to show off about? Or even more dire, if you need to book an urgent flight and your favorite travel site loads a 10MB image proclaiming it’s “Sale Day”; you wouldn’t really be thrilled about it.
Optimizing only for mobile viewing is just one piece of the puzzle. Making sure content is optimized for mobile networks and continuously monitoring for deviations from your benchmarks requires proactive monitoring. Strategically monitoring from real mobile networks should be another essential component of your website performance monitoring strategy.
Anticipate failures and get a good insurance plan
If you’re a smart trekker, you’re always anticipating what might go wrong in the mountains. Be it the weather turning bad, your fitness deteriorating, an accident involving you or your fellow trekkers—you should always be able to bank on affordable emergency services (without risking a cardiac arrest looking at the expenses of your emergency evacuation). A good insurance plan provides you this assurance.
Similarly, anticipating an IT infrastructure failure is what a good site reliability engineer would do. We’ve seen too many incidents where relying on a single vendor backfired; the Dyn DNS failures, and the AWS S3 outages are great examples. Planning for redundancy has become a critical aspect of an SRE’s job.
Being alerted almost instantaneously of failures becomes the critical difference between your website going down or seamlessly switching over to your backup/secondary vendor services. A great example of this is LinkedIn, they use multiple CDNs (up to six), DNS vendors, and other services. See how the LinkedIn SRE team leverages Catchpoint as a critical component of their website performance monitoring strategy.
Reaching the summit is not the end, continuous monitoring is necessary
So often trekkers lose focus as soon as the summit has been reached. A successful trek also requires you to descend back to civilization, alive. A grim reminder of this was a statistic from a few years ago, when most deaths on the Everest summit happened during descent.
In the same vein, as soon as your website meets your performance requirements, it does not mean you stop monitoring it. A good performance engineering team continuously monitors, benchmarks, and acts on website performance metrics. Catchpoint’s synthetic monitoring and real user measurement are powerful tools that when working together, give you the most comprehensive insights on how your site is performing. Our latest ebook details how Catchpoint can help you monitor all that matters and ensure a consistently great experience for your end users.
The marriage between IT and Business must be a stable one
They say the true indicator for a stable marriage is if you could put up with your spouse in the most uncomfortable conditions. Making the best use of situations when things go out of control and work together as a team without the relationship falling to pieces is critical.Similarly, a good indicator of a successful organization, is the cooperation between their IT teams and business teams.
Changes by IT teams and the infrastructures managed by them can have a direct impact on the revenue, while business teams setting unrealistic expectations or cutting back on IT budgets usually bite back on eventual revenues.Setting the right method of evaluation and selecting the right performance metric to monitor becomes a joint exercise. Catchpoint’s Guided Intelligence, helps IT teams gain actionable insights from performance data and helps business owners correlate impact of performance on revenue.
Trekking up the highest mountain range on earth was a truly life-altering experience and something I will never forget. While the climb had me contemplating some of life’s greatest questions, I couldn’t help but also ponder the importance of website performance monitoring to our daily lives. What can I say, once a performance engineer, always a performance engineer.
Until recently, Anand was Solutions Engineer at Catchpoint. He trekked up to Mt. Everest base camp accompanied by his colleague Bharath Kumar, APAC Sales Leader for Catchpoint.