Infra
Nginx 502 Bad Gateway Error When Replacing Containers: Cause and Solution
Encountering Nginx 502 errors when replacing containers? Let's dive into the cause analysis and how to resolve it by configuring `proxy_next_upstream`.
Infra
Cloud Build Cost Skyrockets: How a Single 'machineType' Line Killed My Free Tier
I accidentally disabled Cloud Build's generous free tier and racked up a $36k bill by adding a single `machineType` line to my build configuration. Here's how I fixed it.
Infra
How to Fix Search Engine Indexing Issues Caused by robots.txt Block Errors
Sharing from Infra experience how to solve the problem of being missing from search engine indexing due to robots.txt configuration errors. Check out the solution for unintended path blocking.
Infra
Vertex AI 'Resource exhausted' (429) API Rate Limit on a Single VM
Running a full AI product alone on a small VM comes with unique challenges, including hitting API rate limits unexpectedly. This postmortem details how we diagnosed and resolved a persistent 429 error from Google Cloud's Vertex AI.
Infra
Shrinking a Node.js Docker Image from 2.5GB to 300MB: Leveraging standalone server.js
The Node.js Docker image size problem, secrets to shrinking it from 2.5GB to 300MB! Introducing how to use standalone server.js and pkg.
Infra
APScheduler's Advisory Lock Failure: My Solo VM's Scheduler Died Permanently
When your background jobs stop running, and the logs are silent, it's time for a deep dive. I discovered my APScheduler was permanently dead due to a flawed leader election mechanism relying on PostgreSQL advisory locks.
Infra
Disk Full: How `docker image prune` Saved My Solo VM from a Deploy Freeze
A seemingly simple deployment process froze my entire AI product because the VM's disk filled up. Learn how `docker image prune` became a critical part of my automated deployment pipeline.
Infra
GitHub Actions vs. GCE Pull Poller: My Single VM Deployment Battle
Discover how a seemingly minor deployment issue with an empty .next directory led to a critical decision: ditching GitHub Actions for a simpler, more robust Google Compute Engine pull poller on my solo-run AI product.
Infra
Next.js 14: 'Could not find the module in the React Client Manifest' — the real cause nobody tells you
A 500 that only happens in production, only sometimes. The culprit wasn't my code — it was the absolute build path baked into the RSC client manifest. Here's the root cause and the atomic-swap fix.
Infra
My Home Wi-Fi Got Slow — Diagnosing Suspected Unauthorized External Use with Data
My internet has been feeling sluggish. Could someone else be using my Wi-Fi? This is the diagnostic process, narrowing down 'Is it my PC / Is it someone else?' using data from ping, fast.com, and the network tab in Task Manager.
Infra
Heap Cap Exceeded RAM on e2-medium 4GB — Recalculating PM2 Memory Allocation
The combined `--max-old-space-size` of 5 PM2 cluster processes was exceeding the server's RAM. It wouldn't crash during normal operations, but would OOM during spikes. This post details how I reallocated the Node / non-Node budget within the e2-medium 4GB instance.
Infra
The Day `package.json` Disappeared and My Deployment Broke — A Next.js Sibling Build Recovery Story
My GitHub Actions build died with 'Cannot find module browserslist'. The culprit? `package.json` had vanished from the working tree. This is the story of how I recovered from a ghost state where PM2 was alive but new builds kept failing, using `git checkout`.
Infra
Nginx Configuration Change: Isolating /api/me/* Requests to Port 3001 for API Routing Optimization
This post explains how to optimize API routing and enhance service isolation by isolating /api/me/* requests to port 3001 through Nginx configuration changes.
Infra
3 Deployment Pipeline Regressions Fixed with `printf` and `distDir` Isolation!
We tackled heredoc errors in our deployment pipeline, the absence of the `.next` directory, and RSC manifest path issues by implementing `printf` and isolating the `distDir`.
Infra
7 Infra Improvement Strategies to Prevent Next.js Deployment Build Failures in 2026
7 Infra improvement strategies to prevent Next.js deployment build failures in 2026. Solutions for issues like `--preload` logic, extension support, and build unification.
Infra
GitHub Actions Deployment Downtime Due to Import Errors? Solved with a 3-Step Safety Net!
Did your service go down during a GitHub Actions deployment because of an import error? Build a stable deployment environment with this 3-step safety net.
Infra
0 Minutes of Downtime! Mastering Next.js Zero-Downtime Deployments with Atomic Swap
Have you experienced frequent service interruptions during Next.js deployments? This post introduces how to implement 0-minute downtime deployments using the Atomic Swap technique.
Infra
Google AI Studio 503 Errors? Migrated to Vertex AI and Converted 27 Files Instantly!
Experiencing frequent Google AI Studio 503 errors? I migrated to Vertex AI and successfully converted 27 files in an instant! Learn how to ensure stable service operations.
Infra
Blocking Bot Traffic with nginx + fail2ban, No Cloudflare Needed
I ran into issues with Cloudflare conflicting with SSE streaming, causing disconnections. I switched back to running a single GCP instance. Instead, I built my own scanner/bot defense line using just nginx limit_req + fail2ban + sysctl.