ivo's home

ServerOS panel is a web interface for managing Linux servers you already own. It connects via a local daemon to handle deployments, backups, and monitoring without constant terminal use. The setup requires Docker and specific DNS records for subdomains to function correctly.

It is worth trying if you want a unified view of your VPS fleet but need to maintain the underlying infrastructure yourself. The catch is that you must manage the DNS configuration and keep the daemon running on each server. It is not a replacement for cloud providers, as your servers remain on their respective accounts.

more: https://github.com/serveroshq/panel

my other side: ivo.blogspot.com

BuildKit is the container image build toolkit now driving Docker Engine 23.0 and later. It handles concurrent dependency resolution, instruction caching, and multi-platform image generation through a standalone daemon and client interface. The documentation covers setup across Linux, macOS, and Windows, along with advanced cache export options for registries and cloud storage.

If your workflow already uses standard docker build commands, you likely do not need to read this unless you are migrating to a standalone buildkitd daemon. The guide assumes comfort with low-level build infrastructure and LLB operations, making it less useful for casual users who only need basic image compilation. It is a reference for engineers optimizing CI/CD pipelines rather than a general introduction to containerization.

more: https://github.com/moby/buildkit

my other side: ivo.blogspot.com

This tool runs containers in a small Linux VM on macOS, Windows, and Linux. It speaks the Docker Engine API, so your existing docker CLI and Compose setups work immediately. The project claims a 2.5 ms container start time, which is significantly faster than Docker Desktop.

It is worth a look if you are tired of Docker Desktop waking up slowly or eating memory. The binary is lightweight and the daemon stays small. The main caveat is that it is a new project without a built-in builder yet, so you may need external tools for complex image builds.

more: https://github.com/libracoder/kiri

my other side: ivo.blogspot.com

Cal.diy is the MIT-licensed community edition of Cal.com, built on Next.js, Prisma, and Postgres. It strips out SSO, organizations, and other enterprise tools to provide a pure open-source scheduling stack. You run the entire infrastructure on your own hardware without needing a license key or external account.

It is only worth your time if you have specific infrastructure needs that commercial Cal.com cannot meet. The project explicitly warns against production use, citing the lack of official support and the requirement for advanced server administration. If you just need standard scheduling for a team, the commercial version is the safer bet.

more: https://github.com/calcom/cal.diy

my other side: ivo.blogspot.com

Overlook is a single Go container that aggregates status from existing monitoring tools like Prometheus, Proxmox, and UniFi into one read-only dashboard. It runs version 0.5 as a stateless service requiring no database, allowing it to display server health, storage, and app status without modifying your infrastructure. The interface is a fixed layout that pulls data via GET requests only, ensuring it never changes anything.

It is useful if you want a single, low-maintenance view of your homelab without the overhead of configuring a new monitoring stack. The main limitation is that it depends entirely on your existing sources, so if you do not already run tools like Gatus or Uptime Kuma, you must set them up first to derive any value from the dashboard.

more: https://github.com/bnlqn/overlook

my other side: ivo.blogspot.com

This Node.js app runs in a Docker container and serves a single-page interface for your homelab. It polls your services to display real-time status lights and allows you to add custom widgets for dynamic data. Configuration is handled via a single YAML file or a basic UI editor.

It is a solid lightweight alternative to heavy enterprise tools if you just want a clean start page. The main catch is that status checks are simple HTTP hits, so complex application health requires custom scripts. It is not for you if you need deep service dependency mapping.

more: https://github.com/lissy93/dashy

my other side: ivo.blogspot.com

Gatus is a health dashboard that monitors services via HTTP, ICMP, TCP, and DNS queries. It evaluates response times, status codes, and certificate expiration using defined conditions. A single Docker command deploys the stable container to port 8080.

It suits developers who want a lightweight alternative to heavy enterprise monitoring stacks. The alerting list is extensive but the tool remains a single binary. Skip it if you need complex distributed tracing or advanced log aggregation, as it only handles basic endpoint health checks.

more: https://github.com/TwiN/gatus

Changedetection.io is a self-hostable web scraper that flags page changes and routes alerts to Discord or Slack. It features an LLM integration that converts raw HTML diffs into plain-text summaries or filters noise based on natural language prompts.

The Docker setup is straightforward for standard static sites, making it a solid alternative to brittle RSS feeds. However, the LLM filtering requires a separate API key or local model setup, which adds configuration overhead. It is not a headless browser automation tool, so complex dynamic JavaScript rendering remains a manual configuration hurdle.

more: https://github.com/dgtlmoon/changedetection.io

my other side: ivo.blogspot.com

This Go-powered tool manages a small fleet of personal servers using SQLite and systemd. It supports Hysteria2, VLESS, and AmneziaWG while idling at 16 MiB of memory on a 1 vCPU box. The panel automates node setup via SSH and provides a single subscription link for clients like Mihomo.

The project is in beta stage with 64 releases already shipped, so expect breaking changes before version 1.0. It is not for running large corporate networks but works well for family or small friend groups. The interface has Russian defaults that you will need to adjust.

more: https://github.com/Mistgate/mistgate

my other side: ivo.blogspot.com

n8n is a fair-code platform for building AI agents and workflows using a visual canvas and custom code. It connects to over 1500 integrations and supports JavaScript and Python execution. The standard deployment uses Docker and runs locally on port 5678.

It is a solid alternative to closed-source tools if you need full control over your data stack. The Sustainable Use License restricts commercial redistribution, and the interface complexity requires a developer mindset. Skip it if you want a simple, no-code drag-and-drop experience without backend knowledge.

more: https://github.com/n8n-io/n8n

my other side: ivo.blogspot.com