Own product · Taxi dispatch SaaS
TaxiDex
My own multi-tenant taxi dispatch SaaS: API, dispatcher portal, driver and passenger apps, and the servers underneath.
Freelance · nginx
I am one senior engineer who administers Ubuntu servers and nginx for live products, because I run my own. TaxiDex runs on Linux with nginx doing TLS, caching and security headers, and my company site is served by nginx under a strict Content-Security-Policy. I work remotely, by the project or month by month.
The process is down or listens where nginx is not looking. The logs say which before I touch the configuration.
The step-by-step fixA slow upstream outlasts the proxy timeout, and the right fix is usually in the slow part, not the timeout.
The step-by-step fixThe proxy treats an idle WebSocket as finished and closes it.
The step-by-step fixAn intermediate certificate is missing from the chain, so some clients fail while others pass.
The step-by-step fixA renewal that worked for months stops after a change nobody connected to it.
The step-by-step fixBrute-force traffic fills the logs, and the fix is to block it, not just to ignore it.
The step-by-step fixEither. A one-off hardening or migration is a project, and ongoing care is available as a monthly plan.
Ubuntu, which is what I run day to day.
Yes. I take a backup first, build and test the new server, switch over and keep the way back until the live system checks out.
Yes, I build the APIs and apps as well, so a server problem is traced into the code when needed.
The app behind nginx did not answer in time. I wrote a fix page on 504 timeouts: find the slow request instead of only raising the timeout.
Proof
Own product · Taxi dispatch SaaS
My own multi-tenant taxi dispatch SaaS: API, dispatcher portal, driver and passenger apps, and the servers underneath.
Company website · Static generator
My company site: 804 pages from a Python and Jinja generator, with 15 free browser tools.
From the fix library
404 Not Found nginxAngular routes return 404 from nginx when you refresh or open a linkIngress manifests using extensions/v1beta1 are rejected after a Kubernetes upgradeingress-nginx controller stops working after upgrading Kubernetes to 1.22 or later413 Request Entity Too LargeUploads larger than 1 MB fail with 413 behind the Kubernetes NGINX Ingress controllerWebSocket upgrade requests rejected with HTTP 403 (browser developer tools, request to /api/v2/objectManager)The MinIO Console will not work behind an nginx proxy: no login, 403 on the WebSocket, or repeated disconnectsWebSocket connection to 'ws://localhost/_next/webpack-hmr' failed: WebSocket is closed before the connection is establishedWebSocket to /_next/webpack-hmr fails when the Next.js dev server is behind nginx(13: Permission denied) while connecting to upstreamnginx cannot connect to the Gunicorn socket: Permission deniedNext step
A discovery call within 48 hours, and a kickoff typically inside two weeks.