Sentry

If you're headed to @WeAreDevs World Congress Europe next month, join @janpeer on Thursday, July 9 at 2:10 pm to find out why Zero-Trust Isn’t Optional Anymore. See you on Stage 9!

When your CFO looks at the Datadog bill

@AriesTheCoder @HarryStuck77 @droid @zeeg

@HarryStuck77 @droid My bro, just use @sentry. @zeeg - please fix this guy 🙃

@droid Loop every 1 hour: Check the Error log. Categorize the errors and create new milestones. Continue fixing milestones.😍

Join @virtualmilin and @KarimJalbout at RAISE Summit as they discuss the benefits of self-healing software vs human intervention when debugging. #RAISE2026

Fixing React Native Apps: Don't Let Bugs Crop Up x.com/i/broadcasts/1…

When AI generated code causes a production incident, who should take responsibility? One of our staff software engineers, @JoshFerge, can let you know his thoughts 👇

The next #SFRuby meetup is hosted by our friends @sentry 💜 Talks by @marcoroth_ @camertron and our Martian Alexander Baygeldin, plus our new tradition: a full round of 60-second intros: everyone in the room comes on stage. It's super fun! See you there!

AI security is a shared security model between users and vendors @sentry has the right model here where they protect the user from themselves However you can't always count on the vendor to do the right thing and have to protect yourself too.

This could be you, too. Thanks for the love @jhhayashi 🫡

Skimmer's support tickets used to say "my app didn't work today" — from field techs already out cleaning pools. Reproducing the issue: days of back-and-forth. With Sentry Session Replay: minutes. and a ~95–99% reduction in MTTR. Read the full case study: sentry.io/customers/skim…

We're LIVE for our first Sentry 201 event with @NikolovLazar. Tune in here: x.com/i/broadcasts/1…

Sentry 201: Making Sentry More Useful x.com/i/broadcasts/1…

Our 201 series starts today so we can teach you how to make Sentry more useful. Starts in an hour: sentry.io/resources/201-…

In the Railway incident an agent found a stray API token and deleted a prod database in 9 seconds. @zeeg (CPO, @sentry) makes the realist case: the model is rarely what holds agents back, the infrastructure around it is. Access scoped to the task is where this stops.

Full episode: insecureagents.com/episodes/david…

"I'm gonna build another Jira extension and I'm gonna love every minute of it" - @zeeg New podcast just dropped, link below for the full episode

Everyone wants to believe agents are running their codebase overnight but most people are still refining slop. I really enjoyed my conversation with @zeeg > it's the harness, not the model > verification, not code gen, is the unsolved problem > Mythos is FUD > MCP isn't dead

"why not just MCP + Claude?" — this is why. Same question for both: avg cost per conversation across your agent's gen AI calls. Seer: right answer in ~5s. Claude + MCP: several turns, tons of tokens, still wrong. @sentry has the telemetry. Seer will query it right.

he is right, I fear.

It’s a great day for all who love free merch! Time for a drop, so get to work. Find the code in 201, Don’t get it? Go learn…ain’t that fun.