In this episode, we talk about scale and the hard realities of system failure in large tech operations. We explore why rare failures become common at scale, and what it takes to build systems that can handle that pressure. From predictive diagnostics to component redundancy, we share practical insights on keeping high-performance and AI infrastructure resilient. This is not theory, it is grounded in real-world lessons from managing complex environments and learning how to plan, isolate, and adapt when things go wrong.
Hazel walks us through the Hackyderm.io leaving the basement migration. We also talk much more generally about Mastodon fediverse and scaling distributed systems.
This podcast is like a super class in what it takes to scale infrastructure and systems, especially live and under duress. Every minute of this conversation is worth listening to twice.
In the Cloud 2030 podcast episode discussing Hacky Derm’s scaling challenges, Rob Hirschfeld commends Hazel weekly for exploring the intricacies of exponential growth and federated platform integration. He highlights the significance of core architectural design decisions, such as Twitter’s use of immutable IDs for tweets and the necessity for sharing media files in federated systems. Hirschfeld emphasizes the impact of early design choices on an application’s lifecycle, resilience, and scalability, encouraging listeners to delve into the insightful January 31st episode and join the Cloud 2030 community for ongoing discussions at the2030.cloud.
What are the challenges of interconnectedness and transparency, specifically concerning Kubernetes and cloud native applications?
We have a fascinating discussion sparked by the question of how exposed we are. What happens when something we don’t know is connected is open and exposed as hackable? What happens when it closes, and we didn’t know?
We talked about how this is inherent in the architecture of cloud native applications and what you can do about it.
This discussion should get you thinking about how to architect not just your applications, but the platforms that you need to connect together to make them work.
