Memfault ads

When I was an embedded software engineer, RMA’d devices would land on my desk with a sticky note: “doesn’t turn on” or “screen character is corrupted.” No data trail. No context. Just hours of guesswork ahead. I’d pull devices apart, solder wires onto the board, comb through code, and try to reproduce the issue. Sometimes, I’d get lucky, but often the issue was impossible to reproduce. My cofounders and I loved building firmware, but we couldn’t picture going through that process for another decade. We started discovering interesting workarounds like crashing on purpose and drilling into core dumps. We realized that instead of relying on customer recollections or endless trial and error, there were helpful metrics we could use to understand the issues more intelligently. Memfault was born from that insight. We built a platform that can capture a complete snapshot of what happened the moment a device failed. We unified error data, system states, and performance metrics all in one place, plus the OTA infrastructure to fix problems remotely. We went from sketching vague outlines of suspects and hoping for the best to shining a spotlight on the culprit and catching the bug in the act. Our mission is to build better connected devices, but if we can keep a few firmware engineers from staring at a blank screen at 2 a.m., that is a win, too. Learn more about us: https://lnkd.in/eSbPTg6P

Embedded Linux developers rarely stick to one programming language—and that makes building observability harder. Some teams run C or C++, others prefer Go, Rust, or Python. Some even rely on full interpreters for Ruby or JavaScript. That diversity makes crash reporting a challenge. You need visibility across environments that don’t always speak the same language. Here’s how we’ve been closing that gap in the Memfault Linux SDK: • Python backtrace support- Unhandled exceptions are now captured and pushed to the cloud alongside other device issues. • Generic trace protocol- Any language that emits a text backtrace to standard error can now send it to Memfault with a single command. • eBPF integration- Development is underway to track per-process disk writes with finer granularity. • Expanded metrics- Added visibility into disk, memory, ethernet, and EMMC lifetime statistics, available out of the box. If you’re working with Linux-based devices, you can explore the full changelog here: https://lnkd.in/eJi455xM

🎧 This morning, we crossed 3,000 streams and downloads on Coredump Sessions. We’re big fans of live sessions and on-demand video, but we're happy to help the embedded community tune in on the go (we know your hands are busy, often soldering, debugging, or reviewing logs). If you’re new here, Coredump Sessions is where we sit down with the embedded engineers, IoT Leaders, and connected device experts shaping the future of embedded systems and dig into the lessons behind the bugs, breakthroughs, and builds. In our last episode, reMarkable CTO Nico Comier shared how customer jailbreaks actually helped his team build a better product and why they chose to lean into openness instead of locking things down. Catch up on some of our favorite sessions here: https://lnkd.in/d9BPZW4Z

Ultrahuman didn’t stumble into success with their fitness ring. They built it with intent, speed, and a dead-simple goal: earn user trust in a category full of overpromises. In the next Coredump, we sit down with Ultrahuman Co-Founder Vatsal Singhal to unpack how Ultrahuman made bold product bets early and backed them with execution. We’ll dig into the real engineering decisions behind building hardware-software products that work in the real world, not just in a demo. We’ll also talk sensors, AI, feedback loops, and why moving fast doesn’t mean shipping junk. 📅 Wednesday, Oct 29th at 8 AM PT | 11 AM ET | 5 PM CET Register to attend or get the recording: https://lnkd.in/efThzg_F

Stop flying blind after release. Memfault gives you deep observability into your devices in the field. Performance trends, failure patterns, and real-time metrics, all in one place. Book a demo!

Debugging firmware shouldn’t take weeks. With real-time data from your devices, you can go from crash to root cause in minutes. Book a demo!