9.
HN
Upright: An Open Source Synthetic Monitoring System
Upright is an open-source synthetic monitoring system designed specifically to track and manage services such as Basecamp and HEY. It provides a robust solution by offering health checks from various global locations, which notify users when service issues are detected. Unlike existing tools like Pingdom, Upright excels through customizable, authenticated browser tests that ensure precise location control for these checks. The system integrates seamlessly into an open-source observability stack and includes four distinct probe types: Playwright for browser-based testing, HTTP for status code verification, SMTP to confirm email server functionality, and Traceroute for analyzing network paths.
Upright's architecture is built on a Rails engine deployed using Kamal on cost-effective VPS nodes. It leverages SQLite for data storage, Solid Queue for managing jobs, and utilizes Prometheus along with AlertManager for metrics collection and notification management. Additionally, it incorporates OpenTelemetry for tracing capabilities, all while supporting a user-friendly dashboard UI available in both dark and light modes.
Deployment of Upright is straightforward; users can set up the system on VPS nodes from providers like DigitalOcean or Hetzner. The setup involves configuring various probes to run across multiple locations, which helps in distinguishing between regional service disruptions and complete outages. The tool’s availability under the MIT license via RubyGems and GitHub ensures easy access for customization, making it a versatile option for synthetic monitoring requirements.
Keywords: #phi4, AlertManager, DNS Subdomains, DigitalOcean, Grafana, HTTP Probes, Hetzner, Kamal, MIT License, Multi-Site Deployment, Open Source, OpenTelemetry, Playwright Probes, Prometheus, Rails Engine, RubyGems, SMTP Probes, SQLite, Solid Queue, Synthetic Monitoring, Traceroute Probes, Upright, VPS Nodes
dev.37signals.com an hour ago
|
10.
HN
The OpenClawification of the Web
The "OpenClawification of the Web" highlights how OpenClaw, a versatile AI platform, has underscored critical vulnerabilities in current technology infrastructure, particularly concerning security, trust, and autonomy. As users employ OpenClaw to create autonomous agents—comparable to personal digital assistants like "Jarvis"—concerns have arisen regarding significant security risks due to the platform’s extensive permissions and susceptibility to context poisoning. These issues have spurred major tech companies into action, prompting them to develop frameworks aimed at enhancing trust and control over AI interactions with real-world data. Innovations such as Visa's Trusted Agent Protocol and Mastercard's Agent Pay are designed to secure transactions, while sandboxed accounts and audit trails provide better monitoring and restriction of agent activities. The widespread interest in OpenClaw has catalyzed a shift towards establishing more robust infrastructure, facilitating safer integration of AI into everyday life beyond the basic functionalities of advanced chatbots connected to APIs. This represents a pivotal advancement in ensuring secure interactions between AI agents and real-world systems.
Keywords: #phi4, AI agent, Apple Watch, DigitalOcean, GitHub stars, Moltbook, OpenClaw, PicoClaw, Trusted Agent Protocol, Web, integration, orchestration layer, sandboxed accounts, security vulnerabilities
elliotbonneville.com an hour ago
|
139.
HN
Upright: An Open Source Synthetic Monitoring System
Upright is an innovative open-source synthetic monitoring system designed to enhance service reliability across diverse geographic locations. It provides a robust alternative to traditional tools like Pingdom by offering customizable browser checks that are authenticated, along with health assessments through various probe types such as Playwright, HTTP, SMTP, and Traceroute probes. The architecture of Upright is based on a Rails engine, enabling deployment over multiple global sites utilizing VPS nodes managed via Kamal. This system strategically executes probes in different geographic regions to effectively identify outages or localized issues. Metrics are reported through Prometheus and AlertManager for alerts, while Grafana supports data visualization capabilities. Integration with OpenTelemetry enhances tracing and logging functionalities.
Upright is positioned as a cost-effective monitoring solution, capable of being deployed on economical servers such as DigitalOcean or Hetzner, with the total setup potentially costing under $20 per month. It features a straightforward setup process facilitated by Rails generators and offers comprehensive configuration options for local development, multi-site deployment, and alerting systems. The platform is available through RubyGems and GitHub, distributed under the MIT license, emphasizing its commitment to providing users full control and seamless integration into existing open-source observability infrastructures.
Keywords: #phi4, AlertManager, DNS Subdomains, DigitalOcean, Grafana, HTTP Probes, Hetzner, Kamal, MIT License, Multi-Site Deployment, Open Source, OpenTelemetry, Playwright Probes, Prometheus, Rails Engine, RubyGems, SMTP Probes, SQLite, Solid Queue, Synthetic Monitoring, Traceroute Probes, Upright, VPS Nodes
dev.37signals.com 10 hours ago
|
210.
HN
Infrastructure-as-Code is the wrong abstraction
The article critiques Infrastructure-as-Code (IaC) for its complexity and cloud-specific nature, likening it to writing assembly language due to the intricacies involved in managing modern infrastructure components. The author proposes an application abstraction layer that simplifies deploying applications as self-contained units with all dependencies included. A suggested solution is using a tool like Defang, which utilizes Docker Compose files for provisioning cloud resources such as containers, managed databases, and load balancers without requiring Kubernetes or VMs. This approach maintains cloud-agnostic configurations, facilitating application deployment across different providers like AWS, GCP, and DigitalOcean with minimal changes.
Defang employs Pulumi to manage infrastructure, allowing stateful services like PostgreSQL to map to their managed equivalents, offering features such as backups and high availability. It also supports AI model dependencies by mapping them to managed LLM services. The tool's goal is to reduce complexity and vendor lock-in by enabling developers to describe applications once for deployment anywhere, despite some limitations in abstraction ("leaky"). Defang aims to provide reliability and safety through rule-based provisioning while addressing demands for private deployments due to enterprise compliance requirements.
The author seeks feedback on this approach and its potential challenges, particularly the use of Compose files as a basis for cloud infrastructure declaration instead of relying solely on code generation methods like Terraform.
Keywords: #phi4, AWS, Defang, DigitalOcean, DigitalOceanKeywords: Infrastructure-as-Code, Docker Compose, GCP, IAM policies, Infrastructure-as-Code, LLMs, Pulumi, SaaS, Terraform, VPCs, VPS, abstraction, cloud, clusters, compliance, databases, load balancers, managed databases, private deployments, server programming
defang.io a day ago
|
263.
HN
I Built My Mobile Second Brain
This guide provides a comprehensive method for establishing a mobile-accessible "second brain" using a combination of DigitalOcean for hosting, Obsidian for note-taking, Claude Code for artificial intelligence interactions, and Happy CLI for remote mobile control. The process begins with setting up infrastructure on a DigitalOcean droplet at approximately $24 per month, running Ubuntu 24.04 LTS, including configuring SSH access via key authentication and creating a non-root user. System preparation involves updating the system, installing essential dependencies like `xvfb` and `openbox` for virtual display management, along with Node.js and other utilities.
Obsidian is installed from a `.deb` package and set to operate headlessly using `Xvfb`, enabling it to run in a virtual environment. It's configured with Sync to ensure notes are accessible across various devices. The AI component is integrated via Claude Code, deployed on the droplet for interaction within the Obsidian vault, requiring authentication and functional testing.
To facilitate mobile access, Happy CLI is installed, allowing users to control both Obsidian and Claude Code from a mobile device by establishing a secure SSH tunnel between the phone app and the droplet. Systemd services are configured to manage these applications persistently, ensuring they automatically restart on reboots or disconnections. Verification through service status checks, vault file accessibility, and interaction tests between the mobile and droplet systems is crucial for troubleshooting.
Regular updates for system packages and key applications like Node.js, Happy CLI, and Claude Code are recommended to maintain security and functionality. While the setup incurs a monthly cost of $24, users who frequently utilize this system might consider transitioning to a local Raspberry Pi configuration as a cost-effective alternative after about eight months of usage. This approach integrates cloud-based services with personal mobile access, providing robust note management and AI interaction within Obsidian.
Keywords: #phi4, ARM64, Backup, Claude Code, Cloud, Desktop, DigitalOcean, Droplet, Encryption, Flatpak, Happy CLI, Headless, Linux, Maintenance, Mobile Brain, Nodejs, OOM KillerKeywords: DigitalOcean, Obsidian, Phone, Raspberry Pi, SSH, Swap, Sync, Troubleshooting, Ubuntu, VNC, VPS, Vault, systemd, tmux
robdodson.me a day ago
|
338.
HN
Upright: An Open Source Synthetic Monitoring System
Upright is an open-source synthetic monitoring system designed to oversee services like Basecamp and HEY by conducting health checks from multiple geographic locations using a Rails engine deployed with Kamal on cost-effective VPS nodes. It supports four probe types: Playwright for browser-based interactions, HTTP for status code validation, SMTP for email server assessments, and Traceroute for network path analysis. The system integrates seamlessly into an existing observability stack by utilizing Prometheus for metrics collection, AlertManager for alerting, and Grafana for data visualization.
The customizable probes allow users to monitor diverse service aspects, while the multi-site deployment capability differentiates between regional issues and complete outages by executing checks from various locations. Upright's architecture is built on SQLite for storage, Solid Queue for job management, Prometheus for metrics, AlertManager for notifications, and OpenTelemetry for tracing.
Deployment of Upright can be achieved using VPS nodes such as DigitalOcean or Hetzner, with a typical five-site setup incurring approximately $110 per month. To ensure reliability, metrics are sent to three separate Prometheus instances. Setting up the system involves creating a new Rails application, incorporating the Upright gem, executing an installation generator, and configuring the necessary probes. The platform is accessible on RubyGems and GitHub under the MIT license, simplifying initial setup for users looking to implement effective service monitoring.
Keywords: #phi4, AlertManager, DNS Subdomains, DigitalOcean, Grafana, HTTP Probes, Hetzner, Kamal, MIT License, Multi-Site Deployment, Open Source, OpenTelemetry, Playwright Probes, Prometheus, Rails Engine, RubyGems, SMTP Probes, SQLite, Solid Queue, Synthetic Monitoring, Traceroute Probes, Upright, VPS Nodes
dev.37signals.com a day ago
|
368.
HN
I built a free alternative to Datadog Synthetic Monitoring using Playwright
Vajid, founder of a small development agency, created an alternative to Datadog's Synthetic Monitoring service using Playwright, Node.js, and BullMQ. His motivation stemmed from encountering scenarios where websites indicated "200 OK" status despite functional issues, such as JavaScript errors affecting critical user interactions like a broken "Add to Cart" button. To address this, Vajid developed a tool that prioritizes checking specific DOM elements over merely confirming HTTP statuses. The tool functions by launching headless browsers to navigate URLs and verify essential elements' presence, capturing screenshots and console logs if key processes fail. This approach aims to provide more accurate detection of website issues.
While similar tools like Datadog exist, they can be financially burdensome for small startups or independent developers due to their high costs, typically around $15 per check. Vajid's tool is designed not as a competitor but as a "loss leader" to demonstrate his agency’s capabilities to potential enterprise clients. The core service remains free for the community, with Vajid covering infrastructure expenses on DigitalOcean. Additionally, he supports 5-10 student or open-source projects by offering hosting and monitoring credits.
Vajid is actively seeking feedback, particularly concerning the handling of false positives, and is investigating advanced DOM diffing techniques to improve the tool's reliability further.
Keywords: #phi4, BullMQ, DOM diffing, Datadog, DigitalOcean, JavaScript error, Nodejs, Playwright, Synthetic Monitoring, Vajid, dev agency, e-commerce site, false-positive handling, free credits, headless browser, infrastructure, monitoring tool, monitoring tool Keywords: Vajid
news.ycombinator.com a day ago
|
581.
HN
I built a free synthetic monitoring suite (Playwright based)
A developer at a small agency developed an open-source synthetic monitoring tool using Playwright to address deficiencies in traditional uptime monitors, specifically their inability to detect silent JavaScript errors during client checkout flows. The suite includes features such as Checkout Defender for payment process verification, Login Validator for authentication checks from US and EU regions, and API Deep-Check to validate JSON structures. Built with React and Node.js and hosted on DigitalOcean, this tool offers a budget-friendly alternative to costly enterprise solutions like Datadog or New Relic. It allows users to conduct basic audits without any signup requirement and is available for testing at Pingsla's website. The developer encourages user feedback to improve the suite further.
Keywords: #phi4, API Deep-Check, Auth Validation, Basic Audit, Checkout Defender, Checkout Flow, Datadog, Dev Agency, DigitalOcean, Enterprise Tools, Feedback, Headless Browser Checks, JSON Structure, Login Validator, New Relic, Nodejs, Payment Iframe, Playwright, React, Silent JS Error, Synthetic Monitoring, Uptime Monitors
news.ycombinator.com 2 days ago
|
1010.
HN
Show HN: AgentProbe – Validate AI agent endpoints across 8 protocols in one URL
AgentProbe is a multifaceted validation tool designed to assess AI agent endpoints across eight distinct protocols using a unified URL interface. Users can input a URL and instantly determine endpoint support for protocols such as HTTP, MCP, A2A/AP2, x402, OAuth, MCP Apps, HTML, and ERC-8004 by clicking "Validate." The tool provides comprehensive feedback, detailing each protocol layer's status, including detected tools, payment networks, SSL validation, agent card metadata, and AP2 detection. Additionally, AgentProbe incorporates a built-in MCP server that allows for programmable endpoint validation. Developed with Node.js 22 and vanilla JavaScript, it is hosted on the DigitalOcean App Platform, with its source code available at FlowMCP's GitHub repository under mcp-agent-validator. The creator invites feedback on their detection methodology, highlighting the tool's capability to offer a thorough multi-protocol assessment through a single probe interface.
Keywords: #phi4, A2A/AP2, AI agent endpoints, AgentProbe, DigitalOcean, ERC-8004, HTML, HTTP, JavaScript, MCP, Nodejs, OAuth, URL, assessment, classification, detection, feedback, layers, payments, protocols, reachability, reputation, server, validation, x402
agentprobe.xyz 5 days ago
|
1154.
HN
Show HN: IP ranges for 22 cloud providers in 12 formats, updated daily
The "cloud-provider-ip-addresses" project on GitHub offers an open-source dataset comprising IP ranges for 22 cloud providers and several bot crawlers, with daily updates in 21 formats like JSON, CSV, SQL, plain text, merged CIDRs, and configurations suitable for tools such as nginx, Apache, iptables, HAProxy, Caddy, and UFW. The project compiles data from official sources, merges overlapping CIDR blocks, and ensures daily updates without using APIs or external services. It serves as a vital resource for applications needing up-to-date cloud IP ranges, including firewall rules, rate limiting, and bot detection, by simplifying access to this information across various formats. This dataset is hosted in the GitHub repository at [rezmoss/cloud-provider-ip-addresses](https://github.com/rezmoss/cloud-provider-ip-addresses), providing a comprehensive solution for managing cloud-related network configurations.
Keywords: #phi4, AWS, Apache, Atlassian, Azure, CIDRs, CSV, Caddy, Cloudflare, DigitalOcean, Fastly, GCP, GitHub Actions, HAProxy, IP ranges, JSON, Linode, Oracle, SQL, Telegram, UFW, Vultr, Zoom, bot crawlers, bot detection, cloud providers, firewall rules, flat files, iptables, nftables, nginx, open-source dataset, plain text, rate limiting, repo
news.ycombinator.com 6 days ago
|
2205.
HN
OpenClaw (MoltBot) as a Service on DigitalOcean
OpenClaw (MoltBot) is an open‑source framework that lets developers build personal AI assistants capable of integrating with messaging platforms such as Telegram, Slack, and Discord. DigitalOcean’s App Platform now hosts a managed “OpenClaw as a Service” offering that removes the need for infrastructure ownership while preserving full code‑defined control over agent behavior, model selection, and channel configuration. The platform automatically provisions container runtimes, networking, observability, and supports zero‑downtime Git‑driven image upgrades. Multiple agents can be defined in a single App Platform specification, with individual resizing or upgrading performed without service interruption, allowing smooth scaling from a single assistant to a fleet of specialized agents. Agents run as background workers behind a private Tailscale network or the DO CLI, with no public URL, and are deployed in disposable, hardened containers that start fresh on each deploy, minimizing drift and patching requirements. Persistent state—configuration, sessions, and memory—is preserved across restarts through optional real‑time backups to DigitalOcean Spaces, keeping the runtime stateless. Two secure production modes are offered: a Tailscale‑enabled Web UI that disables public access, and a headless gateway mode with no inbound ports, both capable of synchronizing state to Spaces. Deployments can be launched via a one‑click Droplet for experimentation or through App Platform for elastic scaling, simplified operations, and predictable instance‑based pricing, enabling teams to expand from one agent to many without managing infrastructure.
Keywords: #gpt-oss:20b, 1-Click Deploy, AI, App Platform, CLI, DigitalOcean, Droplet, OpenClaw, Spaces, Tailscale, VM-based, assistants, elastic scaling, predictable costs, scaling, simple operations
www.digitalocean.com 13 days ago
|
2596.
HN
Show HN: Prvue – Self-managed preview environments for back end apps
Prvue is an open‑source solution that creates isolated preview environments for every pull request, offering the same rapid, on‑demand linking that front‑end platforms like Netlify and Vercel provide but for stateful back‑end services; built with Docker, Terraform, and Ansible, it supports a wide array of frameworks—including NestJS, Laravel, Node.js, Rust, Go, and PHP—on DigitalOcean, and was conceived after the author discovered a similar feature in Railway, aiming to simplify back‑end previewing, with full documentation on docs.prvue.dev, community collaboration encouraged, and demo videos forthcoming.
Keywords: #gpt-oss:20b-cloud, Ansible, DigitalOcean, Docker, Laravel, NestJS, Netlify, Rust, Terraform, Vercel, backend apps, open-source, preview environments
news.ycombinator.com 14 days ago
|
2597.
HN
Show HN: Prvue – Self-managed preview environments for back end apps
Prvue is an open‑source solution that automatically spins up isolated Docker stacks—comprising both application and database layers—on DigitalOcean for every GitHub pull request of a backend project, enabling preview environments similar to front‑end services like Netlify and Vercel. The tool leverages Terraform to provision droplets and Ansible to configure Docker, nginx, and an orchestrator that routes preview traffic and handles teardown upon PR closure or when a configured TTL expires. It supports a range of backend frameworks—including NestJS, Laravel, Node.js, Rust, Go, PHP, and Python—through built‑in templates or user‑supplied `docker‑compose.preview.yml` files, while the Prvue CLI offers commands such as `preview init`, `preview setup`, `preview sync`, `preview status`, and `preview destroy` to manage the environment lifecycle. Users begin with `preview setup` to install the CLI and authenticate, add a `preview-config.yml` (and optionally a `docker-compose.preview.yml`) to their repository root, and then open a pull request; the orchestrator automatically builds the stack and posts the preview URL as a PR comment, with documentation hosted at docs.prvue.dev and forthcoming demo videos, inviting community collaboration and contribution.
Keywords: #gpt-oss:20b-cloud, Ansible, CLI, DigitalOcean, Docker, Go, Laravel, NestJS, Netlify, Nodejs, Python, Quickstart, Rust, Terraform, Vercel, backend, database, environments, isolation, nginx, preview
docs.prvue.dev 14 days ago
|
