Uber Cuts Incidents by 71%, Netflix's Hits 99.999% Uptime

Today, we dive into the impressive technological achievements of market leaders. Uber has successfully reduced incidents by 71% per 1,000 code changes. Meanwhile, Netflix's Pushy server now boasts a 99.999% uptime while managing hundreds of millions of concurrent connections. Impressing, right?

Shifting gears, we analyze the current state of the SaaS market, highlighting the importance of using AI to improve existing service rather than replacing them. We also examine AWS's strategic partnership with Oracle, aiming to deliver a secure and cost-effective database solution tailored for AI applications and of course, we cover the newest OpenAI release.

Also In The Queue

• CloudFlare securing API

• 8 Mile Method

• Changes in Microsoft after CrowdStrike incident

• Do you need separate Node.js backend?

Reduce Friction

Shifting End-to-End Testing Left on Microservices

Uber reduced incidents by 71% per 1,000 code changes by shifting end-to-end testing left. Their strategy focuses on early detection, ensuring issues are caught before they hit production. Using sandbox environments and modular test frameworks, they’ve made testing faster and more reliable.

Taking Steps That Drive Resiliency and Security for Windows Customers

Microsoft recently convened a summit with top security vendors and government officials to tackle modern security challenges affecting Windows customers. Sparked by a significant incident with CrowdStrike in July, the summit aimed to enhance resiliency and protection across the ecosystem. 

SaaS Is Dead, Long Live SaaS!

The SaaS industry has seen major volatility in 2024, with concerns about commoditization, pricing pressures, and competition driven by AI. While some predict disruption, the future remains bright for companies that harness AI to add value and differentiate.

The 8 Mile Method

The "8 Mile Method" takes a cue from Eminem's strategy in the film "8 Mile," where he preempts his opponent's criticisms by addressing them first. Applied to leadership, this means proactively acknowledging potential objections to your strategies or proposals before others raise them.

Being an Engineering Manager at Amazon

Ever wondered how Amazon's engineering managers drive performance at scale? Gilad Naor shares three powerful leadership insights: solve problems with sustainable mechanisms rather than relying on good intentions; eliminate vague language to foster precise, actionable communication; and empower teams with full autonomy to boost innovation and ownership.

Deepen Your Expertise

Next.js vs Node.js 

Since Next.js has become a powerful full-stack framework, building a separate backend with Node.js became an option, not a necessity. While the technologies aren’t competing, each has its strengths depending on the project’s needs. Explore when you can rely on Next.js alone and when a Node.js backend becomes essential.

Master Difficult User Authentication Requirements with OAuth

Implementing OAuth 2.0 provides a flexible framework that supports various authentication methods—such as passwords, passkeys, and third-party identity providers—without requiring changes to your applications. 

Pushy to the Limit: Evolving Netflix’s WebSocket Proxy for the Future

Netflix has evolved its WebSocket server, Pushy, to handle hundreds of millions of concurrent connections with 99.999% reliability. By optimizing scaling strategies, automating components, and introducing direct push and device-to-device messaging, Pushy now supports real-time interactions and cross-device communication.

Protecting APIs From Abuse Using Sequence Learning and Variable Order Markov Chains

Attackers are increasingly exploiting APIs by making sequences of calls that resemble normal user behavior, making them difficult to detect with traditional security measures. Cloudflare has faced it by using variable order Markov chains to model and analyze API request sequences.

Increase Scalability

OpenAI Releases o1, Its First Model With ‘Reasoning’ Abilities

OpenAI has unveiled o1, a new AI model with advanced reasoning abilities that surpass previous models in solving complex tasks like coding and advanced mathematics. While it's more expensive and slower than GPT-4, o1 represents a step toward more human-like AI capabilities.

Stop Trying To Replace Your SaaS Products With AI

Replacing SaaS products with in-house AI solutions is often a misguided and costly move. Rebuilding SaaS often means overpaying on headcount and resources for a subpar product that can’t keep up with SaaS providers' rapid innovation. Instead of falling into this trap, focus on leveraging AI to improve your core product, not replacing tools that already work well.

Why Cloud Migrations Fail

Cloud migrations fail when security responsibilities are unclear, compliance is overlooked, or post-migration management falters. Without strong governance, these missteps lead to costly overruns and risks. Proactive oversight can be the difference between success and disaster.

How Meta Is Reinforcing its Global Network for AI Traffic

Meta faced a major challenge in 2022 as AI-driven traffic surged, putting immense pressure on its global network. To keep up with growing demands, Meta’s networking team has been optimizing data flow and developing new strategies, including caching and advanced traffic scheduling, to prevent traffic overload.

Should We Decompose Our Monolith?

While microservices were all the rage in the 2010s, many companies are now rethinking that approach due to the operational overhead and complexity involved. Shifting from a monolith to microservices—or vice versa—is a significant, but often irreversible decision. Carefully assess unique business and engineering constraints before making such a pivotal architectural choice, especially as trends show a renewed appreciation for the simplicity of monolithic architectures.

Just Cool

AWS Teaming Up With Oracle

Amazon Web Services (AWS) has announced a strategic partnership with Oracle to launch Oracle Database@AWS, aimed at providing unique value to customers. AWS CEO Matt Garman emphasized that the focus of this partnership is on delivering a secure, cost-effective infrastructure that supports a range of AI capabilities.

Let’s stay in touch! 📨 

Do you have any comments about this newsletter issue or questions you want to ask? Drop me a message or book a meeting.