FastAPI Extension for VS Code Now Available

Many of my connections here on LI are Python devs, some of them are using FastAPI - Savannah Ostrowski cooked something especially for you! FastAPI VSCode extension. It has many features already! It will be quite great in future if there is way to send request for manual testing purposes? 👀 https://lnkd.in/dexS972B

𝐁𝐮𝐢𝐥𝐝𝐢𝐧𝐠 𝐚 𝐅𝐢𝐥𝐞 𝐒𝐡𝐚𝐫𝐢𝐧𝐠 𝐖𝐞𝐛 𝐀𝐩𝐩 𝐅𝐞𝐚𝐭𝐮𝐫𝐞 𝟏: Smart Dashboard While building 𝐦𝐲 𝐅𝐢𝐥𝐞 𝐒𝐡𝐚𝐫𝐢𝐧𝐠 𝐖𝐞𝐛 𝐀𝐩𝐩, the first thing I focused on was creating a clean and informative dashboard. A dashboard is important because it gives users a quick overview of their entire storage activity in one place. So I designed a dashboard that shows key insights such as:  𝐓𝐨𝐭𝐚𝐥 𝐅𝐢𝐥𝐞𝐬 – Displays the total number of files uploaded by the user. 𝐒𝐭𝐨𝐫𝐚𝐠𝐞 𝐔𝐬𝐞𝐝 – Shows how much storage the user has consumed so far, helping them track their usage. 𝐏𝐮𝐛𝐥𝐢𝐜 𝐅𝐢𝐥𝐞𝐬 – Indicates how many files are currently shared publicly and accessible through links. 𝐏𝐫𝐢𝐯𝐚𝐭𝐞 𝐅𝐢𝐥𝐞𝐬 – Displays the files that are only visible to the user for personal storage. To make the platform more user-friendly, I also added a 𝐐𝐮𝐢𝐜𝐤 𝐀𝐜𝐭𝐢𝐨𝐧𝐬 𝐬𝐞𝐜𝐭𝐢𝐨𝐧 where users can instantly perform important tasks like: ▫️Upload new files ▫️View all uploaded files ▫️Manage their subscription plan ▫️Check transaction history The goal was to make the dashboard simple, organized, and efficient, so users can manage their files without confusion. This was the first step toward building a complete cloud file management system. #WebDevelopment #FullStackDeveloper #ReactJS #BuildInPublic #DashboardDesign

𝗜 𝗕𝗎𝗶𝗹𝗍 𝗔 𝗖𝗹𝗼𝗎𝗱 𝗖𝗼𝗱𝗲 𝗘𝗱𝗶𝗍𝗼𝗿 𝗜𝗻 𝟴𝘁𝗵 𝗚𝗿𝗮𝗱𝗲 I'm in 8th grade. I built a browser-based code editor called Nestd. It saves your work to the cloud. You can use it from any browser on any device. I built Nestd because I code every day after school. But at school, I had no way to code. No setup, no shared folder. So I decided to build my own solution. Nestd has a VS Code-like editor powered by Monaco. I built it with vanilla HTML, CSS, and JS. You can try Nestd here: https://nest-d.vercel.app/ I'm proud of this project. I'd love to hear your feedback. Source: https://lnkd.in/gF8Ry-xD

A few months ago, SMMUN's registration system was a single FastAPI script on Railway. It worked - until it didn't. I've been studying for my GCP ACE certification, so I took the chance to put it into practice and rewrote the whole thing from scratch. The core idea: don't make users wait for things they don't need to wait for. The API hosted on Cloud Run validates input, uploads the file, and writes to Firestore - then immediately redirects. Everything else (Google Sheets, confirmation emails) happens asynchronously through Pub/Sub and a worker service running in Cloud Run functions. The interesting problems were in the details: - Pub/Sub can deliver the same message more than once. So the worker uses Firestore transactions to claim submissions atomically - if it's already been processed, it exits safely. - Files are stored privately in Cloud Storage and accessed through signed URLs, so nothing is ever publicly exposed. - Every request gets a request_id that propagates from the API through Pub/Sub to the worker, which makes debugging across services actually possible. The whole infra is managed with Terraform. Open to feedback from anyone who's built similar systems. ↓ Repo in the comments.

Did a little goofing off (?!) this afternoon setting up PostHog to track Ghost page views. Add a read only scope restricted key, a materialized endpoint, and I’ve got something that looks suspiciously like a quick no-extra infra (er… except PostHog) way to display my most read/trending content for site visitors. Or maybe I’ll spin up a little cloud function glue and tag the “trending” posts in Ghost, and skip the extra JavaScript on the front end. Hmm… choices.

This is what I call performance! After the latest refactoring (which uses cloud storage instead of web-server or API to deliver the website artefacts), the full site send.sgraph.ai/en-gb is now loading in 0.181 seconds :) Powerful stats: - 12 languages and 17 countries/cultures - 30 requests loaded - no 3rd party scripts files or dependencies - zero cookies - no user tracking - no JS frameworks (it's all pure Html, JS and CSS) Actually that 0.181s load is with the cache disabled, without that (i.e. normal browsing) it loads in 0.068s (i.e. 68 milliseconds)

What I Learned While Hosting My First Website Using AWS S3. As part of my journey into Cloud Engineering, I recently built a project where I hosted a static website using Amazon S3 Static Website Hosting. At first, it might seem like a simple task, but this project helped me understand an important concept in cloud computing — how websites can be hosted without traditional servers. # Why Do We Need This? In traditional setups, hosting a website usually requires managing a server (like Apache or Nginx). But in the cloud, services like Amazon S3 allow us to host static websites directly from object storage. This approach provides: ✔ High availability ✔ Scalability ✔ No server management ✔ Cost-efficient hosting # How I Implemented It To build this project, I followed these steps: 1️⃣ Created an Amazon S3 bucket 2️⃣ Disabled Block Public Access to allow website access 3️⃣ Enabled Static Website Hosting in the bucket properties 4️⃣ Uploaded my website files (HTML & CSS) 5️⃣ Configured a Bucket Policy to allow public read access 6️⃣ Accessed the website using the S3 static website endpoint # What I Learned Through this project, I gained hands-on experience with: • How Amazon S3 stores and serves static content • How bucket policies control public access • The concept of static website hosting in the cloud • Basic cloud architecture for serving web content # Key Insight This project showed me how cloud platforms simplify infrastructure. Instead of managing servers, we can use managed services like S3 to host websites quickly and reliably. # Project Repository https://lnkd.in/gkBDBPPe I’m continuing to build more projects as I learn cloud technologies and explore how modern infrastructure works. 💬 I would really appreciate any feedback on this project. Also, if you have suggestions for AWS projects that can help me strengthen my cloud skills, I’d love to learn from your recommendations.

I run 5+ Cursor cloud agents in parallel. Keeping track of them was chaos. So I built a Control Tower! 🗼 One terminal command and you can get a view of all your agents on one screen: Running, done, errored - all visible at a glance. You can also launch new agents, send follow-ups, stop runaways. All from the keyboard. How to get it? Just write: ``` npx agents-control-tower ``` That's it. 10 seconds to set up. Repo & Details here: https://lnkd.in/dEZPFuDJ #CursorAI #DevTools #OpenSource #CLI #TypeScript

TIL the X- prefix for custom HTTP headers was deprecated in 2012 🫢 RFC 6648 formally recommends against using X- for new custom headers. The reasoning: if a custom header later becomes standardized, you end up with two names for the same thing. X-Frame-Options is the classic example of this going wrong. But here's the thing: AWS still uses X-Amz-*, Google Cloud uses X-Goog-*, and Azure uses X-Ms-*. Meanwhile, Stripe, GitHub, and Shopify have moved to unprefixed company-namespaced headers (Stripe-Version, GitHub-Media-Type, Shopify-Shop-Domain). The industry is genuinely split. Legacy APIs kept X- because removing it would be a breaking change. Newer APIs dropped it because the RFC says to. The practical takeaway for anyone designing headers today: → New headers: skip the X- prefix. Use a descriptive, company-namespaced name instead (e.g., Acme-Request-ID) → Existing X- headers: leave them alone. The RFC explicitly doesn't recommend migrating existing ones → Consistency within your API matters, but don't let it trap you into perpetuating a deprecated convention forever The RFC uses "SHOULD NOT" (RFC 2119), which means strong discouragement, not prohibition. So you're not breaking any rules either way. But the direction is clear. References in the comments 👇

> "RFC 6648 formally recommends against using X- for new custom headers. Why? Don't encode status... The takeaway is (this) applies anywhere you're naming things in a schema that might evolve: don't encode lifecycle stage into the name. Use namespacing for ownership (who defined this) rather than status (how official is this)." Think ACME-PROPERTY not X-PROPERTY. Definitely somee OpenAPI should have a rethink on.