Dash0

If you’re in cybersecurity like CrowdStrike, you have 1 or 2 minutes, not 1 or 2 days, to get instrumentation in place to fight a threat. Here’s why this catch-22 led to millions of Windows systems crashing: Windows does not want any driver to take down the system. They have a BIG certification process to prevent this. Every software release that runs in the kernel needs to get checked. But if a new cyber threat appears today, you can’t go through that process. Even though CrowdStrike’s core sensor has been tested & verified on Microsoft systems, this new “rapid response” update looked different than previous templates (according to the #CrowdStrike PIR). EX: Windows computers in airports were affected, because they were: 1) online during the push 2) had previously received templates without issue Fabian Lange pointed out this dilemma: You want critical systems like airports, hospitals protected from attackers *immediately.* But this rapid need for protection = rapid software updates = undetected glitches that cause a massive outage like last week’s. One solution = Canary deployments. CrowdStrike now says it will stagger rollouts. This could give customers a say in when they get the push on their machines. Last 10%? First 1%? Then give it 15, 20 minutes to see if it’s working. If not, stop and revert. By the time CrowdStrike did this, it was too late. Millions of screens flashed blue – and we’re still watching the fallout today. CrowdStrike now knows exactly what went wrong. They’ll make sure it doesn’t happen again. We break it down in our new, quick BONUS episode of Code RED (link in the comments 🔗). ➡️ Follow along with me to: - Avoid your next outage - Learn what’s next for observability - Hear from global leaders in the industry about what they’re building

I feel your pain. If you’re a developer in a major code red 🚨 outage, it hurts. Physically. You feel it in your gut. When everyone’s looking at you, and you have no idea what’s wrong… the pressure is on. I’ve felt it myself countless times. So has Geeta Schmidt. It’s why we decided to build observability tools in the first place. To ease the pain for our peers. We recently talked about how she built Humio – which pioneered index-free storage – and how her teams have solved problems involving astronomical amounts of data, 24/7 systems and security operations that need solutions in seconds ⏰. It’s all about preventing the pain. We’ve felt it. You’ve felt it. Let’s get rid of it, huh? Listen to Geeta and I’s full Code RED conversation at the link in the comments 🔗. ➡️ Follow along with me to: - Avoid your next outage - Learn what’s next for observability - Hear from global leaders in the industry about what they’re building

In 5 years, Humio went from 1 crazy idea of index-free storage to a $400M acquisition from CrowdStrike. Here’s how they transformed the observability industry: Before 2015, “observability” was barely a term. Developers were throwing things at the wall, and seeing what stuck. But Geeta Schmidt, Humio’s co-founder, saw a growing need to understand and track the health of your own service, in real time. Like when she helped build Denmark’s national system for medicines. It needed to run 24/7. It could NOT go down. Mission-critical. But her monitoring tools back then were: – Bad at dealing with large amounts of data – Slow to query – Inefficient The solution? Get rid of the indexes. Now, real-time observability is possible. Data on the fly. In an age of astronomical amounts of data. The idea sounded crazy to me at the time. Now it’s the norm. Geeta and I are passionate about building and supporting tools like this (that make your job easier). Listen to our full Code RED conversation at the link in the comments 🔗. ➡️ Follow along with me to: - Avoid your next outage - Learn what’s next for observability - Hear from global leaders in the industry about what they’re building

If you’re a developer in a major code red 🚨 outage, you’re getting alerts from 4 or 5 monitoring tools at once. Which one do you action on? Today, we can make better sense of the noise. How? Observability is constantly evolving. Today, this means new approaches (like AI tools and LLMs) that shortcut your troubleshooting, detect patterns quickly and help you avoid headaches. Jonah Kowall tells me he’s watching 3 key developments: 1. Simplification of observability tools 2. Accessibility to best practices for OpenTelemetry configuration 3. The use of machine learning for better time series analytics The goal? Take observability to the next level. Compare more metrics at once. Ask questions like: Where's the anomaly? Why is this anomaly happening? When you’re stressed, when your system’s failing and everybody's looking at you… We want to help you fix that issue. And find answers fast. Today, we can quiet some of the noise. Listen to Jonah and I’s full Code RED conversation at the link in the comments 🔗. ➡️ Follow along with me to: - Avoid your next outage - Learn what’s next for observability - Hear from global leaders in the industry about what they’re building