The problem of invisibility!
The problem of invisibility!
Date: July 19, 2024.
It appeared to be a typical day.
Hustlers were hustling, and bustling environments worldwide were bustling with activities. Daily life seemed perfect and normal. But something catastrophic was going to happen across the world.
The world was about to witness "the largest" global IT outage in human history.
What seemed to be a regular patch update turned out as "the most faulty" update and invited the "blue screen of death," impacting 8.5 million Microsoft Windows devices.
Digital public infrastructure crumbled!
Businesses, industries, and governments were disrupted!
Thousands of flights were grounded!
Critical surgeries were postponed!
And users of specific banking platforms faced difficulties accessing their accounts!
What caused all of these?
CrowdStrike, a renowned security software company, protected these devices.
CrowdStrike runs "Falcon," a software agent that operates at "ring 0" or kernel level of the OS for end-point detection and prevention. An update to this agent was rolled out with "no regression tests". There was no visibility to the problem, and there was no visibility to the impact in the beginning.
The result of the outage was catastrophic.
It proved one thing - "Software runs the world."
And it underlines one crucial thing in "testing" and "monitoring" space.
No matter how great your code is and your software product is, if your software is not tested in the way the end users use it, there is no guarantee that it will work the way it should. Visibility to the availability of the software is as important as the existence of the software itself.
That is the reason, Observability is a key area businesses worldwide should really focus on.
No matter whether the business is small or large, the emphasis on proper testing and monitoring is crucial. And visibility to the software system to see that it works fine from user locations is more relevent, even in this modern era of AI driven world.
Monitoring the availability of the software is so critical to save humankind from these kind of impactful outages.
Ensuring continuous monitoring will prevent such outages to a great extent.
If "availability monitoring" was in place, the range of disruption these software outages brought in could have been prevented or alleviated.
This is specifically relevant for SaaS and AI Agentic software.
This can happen to any software company and any software product. The only way to prevent such outages is to monitor for outages proactively and take action before they can have a widespread impact.
That's the story of visibility, the problem of invisibility.
For your SaaS offerings, use synthetic monitoring by mimicking critical browser flows using the Microsoft-backed open-source framework - Playwright.
To run these scripts on Amazon Web Services (AWS) worldwide, use PerfAgents. You don't have to live in the fear of unpredictability and invisibility.
Happy New Year!