Systems | Development | Analytics | API | Testing

Do you know where your data is? The number of people who can pat their server and say fondly, “Right here!” is decreasing. Instead, more people are lifting their eyes to the heavens and answering, “Um… up there… somewhere…” McKinsey reports that in 2025, large enterprises have 60% of their environment in the cloud. If you’re considering moving your data assets, processes, and applications to the cloud, you’re in good company.

If you’re on a platform team running Apache Kafka, you know it’s rarely simple. You’re expected to keep it stable, performant, and secure while juggling requests from every direction. Supporting multiple teams and partners leads to operational complexity that never really goes away.

The recent global IT outage experienced by a major cloud hyperscaler was a disruptive, real-world reminder that downtime and service disruptions are inevitable. The event impacted services across banking, retail, and healthcare, and served as a powerful warning that relying on any single provider, or even a single cloud region, creates a critical business vulnerability. This outage highlights the critical risk of a single-provider strategy, rather than an inherent problem with the cloud.

For several hours in the day, millions of users were staring at loading screens or frozen apps. Communication channels went quiet. Games, financial platforms, healthcare apps – and even beds – were all down. The culprit this time: an outage in Amazon Web Services' us-east-1 region. AWS is the dominant provider of cloud compute globally, and us-east-1 is its oldest and busiest region. Countless businesses globally depend day-to-day on services hosted in that region.

During this week’s AWS us-east-1 outage, Ably maintained full service continuity with no customer impact. This was our multi-region architecture working exactly as designed; error rates were negligibly low and unchanged throughout. Any additional round trip latency was limited to 12ms, which is below the typical variance in any client-to-endpoint connection, and well below our 40–50ms global median; this is imperceptible to users and below monitoring thresholds.

In September 2025, we brought AI Engineer to Paris for a two-day conference with 700+ attendees, 25 sponsors, over 47 talks cross 5 tracks. This is the third post in a 4-part series on the event: If you missed AIE Paris, you’ll be able to watch the replays on the Koyeb YouTube channel. Main Stage Day 1 Main Stage Day 2 If you’re not familiar, the team behind the AI Engineer has been running AI Engineer conferences in the US for the past three years.