Rootly

Traditional reliability is built on cause and effect: same code, same infra, same result. Machine learning breaks that rule. You can change nothing and your model still fails. Maria Vechtomova, MLOps Tech Lead, joins Sylvain Kalache, Head of Rootly AI Labs, to unpack why ML systems fail differently from traditional ones and what that means for reliability engineering. Read all of Maria's insights in our latest blog post (linked below).

When operating at scale, figuring out who should get paged isn’t trivial. Especially when teams manage multiple escalation policies or shared services. Rootly makes it simple to route incoming alerts to the right person while giving managers a bird’s-eye view of how paging is distributed across the organization. Learn more in our changelog (linked below).

Attending KubeCon in Atlanta? Join Rootly, Checkly, AuthZed, and more, high above Midtown for a relaxed happy hour at Polaris, the iconic revolving restaurant with panoramic views of the Atlanta skyline. Rotate through 360° city vistas while enjoying greats conversations with SREs, platform engineers, and observability leaders. Expect great drinks, easy networking, and just enough buzz to spark collaboration before the showroom floors open the next day. Register using the link in the comments below.

Incident response, meet AI assistance. 👋 Rootly’s AI helps engineers investigate, communicate, learn from, and resolve incidents faster — automating root cause analysis, generating clear updates, pulling data from across tools, and coordinating the right individuals, all within Slack. That means fewer dashboards, less context switching, and faster fixes. 🪄 Slack 🤝 Rootly = Agentic collaboration 👉 https://sforce.co/48Fd3Ru

Some incidents call for more hands on deck. With Rootly, you can now page multiple responders at once, directly from Slack or the web. Escalate across on-call schedules, teams, or services to get the support you need with fewer clicks and faster coordination. Read more in our changelog (linked below).

We talk a lot about technical skills in incident response: systems design, observability, capacity management. But one of the easiest skills to overlook is knowing when to ask for help. In his latest article, Brandon Chalk (ex-Google) introduces a simple but powerful idea: the Triage Shot Clock, a framework for setting personal time limits during triage so you know when to bring others in. Read the full article, linked below.