Operations | Monitoring | ITSM | DevOps | Cloud

If you’re like most developers, you know that writing code isn’t the bottleneck anymore. AI has made it faster than ever, and chances are you’re already using it. Yet, delivering software is still complex because of everything else you have to manage: fixing vulnerabilities, reducing tech debt, cleaning up feature flags, ensuring test coverage, writing documentation, and the list goes on. That’s why we built Rovo Dev, a context-aware AI agent for developers.

This post is a deep dive into how we improved the P95 latency of an API endpoint from 5s to 0.3s using a niche little computer science trick called a bloom filter. We’ll cover why the endpoint was slow, the options we considered to make it fast and how we decided between them, and how it all works under the hood.

Are your deployments getting stuck waiting for approvals? Your code is ready. Your tests are green. But your ServiceNow change ticket is still holding up the release. In most organizations, this isn’t a people problem or a process problem. It’s an evidence problem. Every release has to prove that it met the required checks — tests, scans, reviews, and approvals. But when that proof isn’t instantly available, everything slows down.

Deploying on AWS efficiently is key. See the updated guide on the best AWS deployment options, covering new features and services.

An artifact repository like JFrog Artifactory is a cornerstone of modern DevOps. It stores binaries, versions, and release bundles — your complete “what.” But when audits or incidents happen, the question quickly shifts from what to how: “How did this artifact get here — and can we trust it?” If all you have is a warehouse of files, you’re left scrambling to reconstruct the story. You check pipeline logs. You pull test results. You cross-reference approvals.

Scaling your application shouldn’t feel like open-heart surgery. It should feel like flipping a switch. Watch your environment adapt in real time. Horizontal scaling. Vertical scaling. One command. Done. You do not want another war room. You want a clear way to add capacity when traffic increases, without editing and testing complex YAML files for hours or manually rolling out scripts across clusters.

In October, technologists from across the financial industry gathered in New York for OSFF 2025 where the general theme was clear: open collaboration has moved from promises to proof. Projects like Fluxnova and OpenGris showed how institutions can build shared, production-grade infrastructure. The Common Cloud Controls and AI Governance Framework demonstrated that regulatory assurance can be achieved collaboratively, not competitively.

When you run Kubernetes at scale, one of the first challenges is understanding what the cluster is actually doing. Workloads shift around, pods restart for normal reasons, and traffic doesn't always follow the patterns you expect. Having clear signals makes day-to-day operations much easier. That's where kube-prometheus-stack helps. It brings Prometheus, Grafana, Alertmanager, and supporting components together as a single package.

When we first introduced the JFrog AI Catalog, it was our mission to provide the industry with a single system of record for governing the complex landscape of internal, open-source, and external commercial AI models. This foundational step was critical for enterprises to move from uncontrolled innovation to delivering AI with trust and confidence. However, the AI landscape is ever-evolving. The challenge for today’s enterprise is already evolving beyond simply managing a library of known models.