devops

赫尔辛基大学通过 MOOC.fi 免费开放其完整的 DevOps with Kubernetes 课程，包含动手实践练习，面向所有人。

博主介绍开源工具 vps-security-hardening，可用一条命令远程自动化完成 VPS 的 7 层安全加固（SSH 检查、系统更新、禁 root 登录、改端口、fail2ban、UFW 防火墙、Docker 防火墙配置），免去手动加固半小时的繁琐流程。

Show HN: k10s 是一个用 Go 和 Bubble Tea 构建的可点击 Kubernetes 终端 UI，作为 k9s 的现代化替代，提供鼠标操作、全局搜索、延迟 informer 以实现即时启动，并内置具备集群上下文感知能力的 AI 助手。

BotBus 是一款新产品，可让你通过手机管理本地运行的编码代理（如 Codex、Claude Code、OpenCode 等），支持继续任务、处理审批、查看文件与终端，并跨 macOS、Linux、Windows 及 iOS/Android 使用。

NVIDIA 发布 AICR v1.0，提供开源、稳定且可验证的 GPU 加速 Kubernetes 集群配置方案，通过版本锁定的配方和签名验证证据简化跨组件兼容性管理。

GitHub reports that Git activity more than doubled year over year to 473.3B events per month and is rebuilding its Git infrastructure to support agentic, agent-scale software development. The post details architectural challenges such as commit latency bottlenecks, surging write throughput, ref contention, and read fan-out from CI and code scanning.

A DevOps engineer shares a first-person story about discovering that an IT contractor had been stealing and reselling company laptops for over a year, and how the lies were uncovered through inventory checks and Apple repair request records.

Agent prompts, skills, and MCP configurations are increasingly treated as serious infrastructure, requiring a single source of truth, version locking, and drift detection—spawning a new wave of configuration management tooling for AI agents.

Montray is a lightweight open-source tray-icon utility that monitors systemd services and arbitrary command-line checks, using a Go-based server paired with a Rust/Slint desktop UI for notifications and health status across local and remote machines.

GitHub reported an ongoing incident affecting GitHub Actions, with job failures and delays in assigning GitHub-hosted runners across multiple runner configurations. Teams are investigating degraded performance and working to mitigate the impact.

A tutorial on using Podman rootless with docker-compose by exposing Podman's Docker-API-compatible UNIX socket and setting DOCKER_HOST, plus tips on disabling rootful Docker and using podman mount/unshare.

A blog tutorial showing how to build a self-hosted HTTP tunneling service — similar to ngrok or Cloudflare Quick Tunnels — using only OpenSSH reverse forwarding and an nginx proxy with wildcard DNS, Let's Encrypt certificates, and ngx_http_secure_link_module for hashed, expiring access URLs.

DevOps Advisor is an AI agent that diagnoses delivery failures and tailors CI/CD, infrastructure, and reliability guidance to a team's needs, using DORA metrics and SRE practices with persistent stack memory across sessions.

A developer recounts an incident where an AI agent got stuck in a runaway loop, burning thousands of dollars on LLM API calls overnight despite high uptime, and asks the community how they build kill switches to stop runaway agent loops.

A single open-source command-line tool called 'reinstall' that reinstalls operating systems (Windows 11/10/Server, Debian, Ubuntu, Alpine, Rocky, Fedora, OpenWrt) with one command, without needing bootable USB drives or BIOS tweaks, even across OS boundaries and on remote servers without IPMI.

The author argues that Verification Engineering is becoming a key bottleneck and career direction in AI-native organizations, and shares seven practices: using Rust-based toolchains to shorten feedback loops, layering programmatic E2E tests, having agents perform regular real-world inspections, reducing the load on dev/test environments, verifying code after it's merged, verifying that tests can actually catch bugs, and building observability for the verification system itself.

ProcBoss launches as a lightweight production process and infrastructure management platform for running, deploying, and monitoring apps across servers, with support for Node.js, Bun, Deno, and other runtimes, plus a Cloud tier offering centralized monitoring, alerts, and hosted status pages.

A practitioner asks how to maintain and deploy customer-specific code generated by an agent harness, ideally keeping customer variants in sync with the core app without dedicated VMs per customer. The post seeks existing tools or platforms that handle end-to-end deployment of agent-modified, per-tenant codebases.

Gergely Orosz reports that four days after a global Firebase outage that crashed iOS apps with analytics enabled, Google has still not responded or updated its status page, raising serious questions about Google's incident detection and communication practices.

GitHub Security Lab开源了gh-secure工具，一条命令即可为公开仓库开启分支保护、密钥扫描、Dependabot、代码扫描等五项安全防护，并支持在Copilot应用与Copilot CLI中使用。