Together with @WorkOS and @awscloud, we’re hosting a landmark event on how MCPs are reshaping reliability for engineering teams. Kicking things off is a panel of industry leaders:.✦ @AarushSah_ (Head of Evals, @GroqInc).✦ @greggmojica (Co-Founder & CEO, @AlloyAutomation).✦

Read JJ's insights on how to make this a reality:

AI SREs can diagnose production issues in minutes. But the hard part comes next: chasing down service owners, untangling conflicting hot-fixes, and trying to coordinate mitigation at speed. AI SRE needs more than just AI. Rootly CEO JJ Tang makes the case for operational context

RSVP here:

RT @LaurenceLiang1: thanks everyone for joining us at the ML reading group meetup last Thursday at @rootlyhq !. shoutout @dhasandev for pre….

There are plenty of great LLMs out there. The real challenge is knowing which one(s) fits your use case. @Aishwarya_Sri0 shares practical ways to choose: from designing your own evals to adopting an experimental mindset.

We tested the new GPT-5 and GPT-OSS models against all leading models in the market on real SRE workflows. The results:.• gpt-oss-120b set a new open-source benchmark, outperforming every OSS model we’ve tested in code understanding and generation. • GPT-5 delivered strong

RSVP here:

MCPs are helping SREs resolve incidents faster and build more reliable systems by turning siloed tools into connected, intelligent workflows. We’re teaming up with @AWSEvents and @WorkOS for a live event showcasing how MCPs are already powering smarter incident response and

Read Shane's guide here:

When senior leaders jump into an incident without context, they can unintentionally slow things down, override priorities, or create confusion for the incident commander. The result? A messier resolution process. In his latest article, Shane Arsenault breaks down how leadership

Self-healing systems aren’t new. But AI is pushing them into far more complex territory. And with that possibility, new challenges arise, including trust. Duncan Winn, VP of SRE at @zscaler, breaks down how the same guardrails that protect production from human error can also

More in the changelog:

No matter how well you plan, sometimes life happens. If you can’t make your on-call shift (or a part of it), use Request Coverage, available in Slack or the web app. Rootly takes care of the rest: notifying your team and coordinating directly with whoever steps in to cover.

RT @oktadev: We co-hosted Rootly’s AI Security Demo Night at Okta HQ, an evening full of real talk around building safer AI. 🔐. Semona shar….

Post-mortem template + prompts + example:

Post‑incident reports can feel like they’re slowing down the return to “business as usual.” But skipping them risks seeing the same incident twice. We built a post‑mortem template that makes it fast:.• Pre‑filled structure for your retrospective.• Prompts to auto‑generate

Gemini CLI has surpassed 60K stars on GitHub in under two months. @jocarrasqueira, Head of DevRel at Google DeepMind, shares why developers are adding it to their terminals nonstop.

Read all their takeways in our latest blog post:

“AI stands for Angry Intern,” says Aaron Erickson (NVIDIA). It’s fast, ambitious, and constantly pushing to ship (whether it’s ready or not). Platform engineers are on the front line as production requests surge, even from outside technical teams. The question: how do you keep