How bad are python cold starts?🥶 Recently improved our P99s on our endpoints at Sentry by over 2 seconds by fixing them using our own profiling tool. Check out my writeup to see how we debugged! blog.sentry.io/how-we-decreas…

Sentry would 100% not be a functioning product if we hadn't used Sentry to debug countless perf and quality issues along the way.

We use Sentry a lot at Sentry, as it turns out 😂

Super fun speaking at the infra.nyc meetup! thank you Megan Reynolds 👩‍💻 for organizing! what an amazing community!

just realized what i'm doing should be in an electron app. i can either rewrite my codebase to javascript, or try and use pyinstaller to create a runnable executable to bundle within electron 😭

GPT-5 API is pretty much unusable at the moment. I imagine OpenAI's capacity management is very stretched right now

having Claude/GPT write *good* end-to-end playwright tests has been a struggle. this is what's worked decently for me. more prompt engineering to do.

it lives

Next open-source project: AI agent email inboxes. No one should endure what I just did with SendGrid 😂

Coding agent instruction config feels pointless at the moment.

A fun internal Sentry error, but also sort of a milestone! When our "issues" postgres table was created, the times seen column was made as an int32. One of our customers has so many occurrences of an issue it's actually overflowed the column! time for an int64!

👀

🚨 Big news! 🚨 You can now kick off a @Cursor_ai Cloud Agent directly from Seer, our AI debugger Seer runs the root cause analysis → passes all the context → Cursor drafts the fix or PR for you Here's Cody De Arkland with a quick demo: