booting md.-nazmus-sadat.portfolio ...

DevOps Engineer @ xCloud

Hi, I'm Md. Nazmus Sadat.

DevOps Engineer

I automate, harden and keep production web infrastructure running at scale.

md. nazmus sadat — zsh
visitor@portfolio:~$
0+

servers operated

0+

incidents resolved

0%

fleet uptime

0/7

on-call coverage

01.about

I'm a DevOps Engineer who lives in the terminal and automates his way out of it.

I run production web infrastructure for a managed cloud hosting platform — provisioning, deployment pipelines, performance tuning, hardening and incident response across a fleet of Linux VPS nodes serving hundreds of customer workloads.

Day to day that means Nginx and OpenLiteSpeed vhosts, PHP-FPM and Node/PM2 services, Docker workloads, MySQL/MariaDB and Redis, Cloudflare edge config, systemd units and cron pipelines. When something breaks at 2am I find root cause from the logs, not from guesswork.

The part I care about most is durability: a fix that survives a panel regeneration, a config that a teammate can read, and a runbook so the next person doesn't have to relearn it. Toil that repeats gets scripted.

Focus
Linux infrastructure, automation & reliability
Stack
Linux · Nginx/OLS · Docker · MySQL · Redis · Bash
Approach
Investigate → confirm → backup → change → verify
Status
DevOps Engineer @ xCloud
02.skills

The toolkit I reach for under pressure.

Linux & Systems~/systems
  • Ubuntu / Debian
  • systemd units & drop-ins
  • Resource & OOM tuning
  • Kernel / sysctl
  • Process & memory forensics
Web & App Servers~/servers
  • Nginx
  • OpenLiteSpeed / LiteSpeed
  • PHP-FPM pool tuning
  • Node.js · PM2
  • Reverse proxies & upstreams
Containers & Cloud~/cloud
  • Docker & Compose
  • Vultr · Hetzner · DigitalOcean
  • AWS (EC2 · S3 · Route 53)
  • Image builds & registries
  • Provisioning & snapshots
CI/CD & Automation~/automation
  • Git · GitHub Actions
  • Zero-downtime deploys
  • Bash & Python tooling
  • Cron & Supervisor pipelines
  • Config-as-code / includes
Data & Caching~/data
  • MySQL / MariaDB
  • Redis (ACLs · object cache)
  • Backups & restores
  • Migrations at scale
  • Query & cache tuning
Networking & Security~/netsec
  • DNS · SSL/TLS · Let's Encrypt
  • Cloudflare (proxy · origin certs)
  • UFW · fail2ban · WAF rules
  • tcpdump / traceroute / mtr
  • Malware IR & hardening
Root-cause analysisIncident commandRunbooks & documentationClear written commsAutomation mindsetPost-incident review
03.experience

Where I've kept the lights on.

DevOps Engineer @ xCloud

June 2025 — Present
  • Operate and troubleshoot a multi-tenant fleet of Linux VPS nodes (Nginx & OpenLiteSpeed) running hundreds of production customer workloads.
  • Own SEV-1 incident response end to end: PHP-FPM 502s, MySQL OOM crash-loops, TLS/origin-cert failures, cache and permission faults — diagnosis from logs to durable fix.
  • Harden services against recurrence — systemd OOM protection drop-ins, right-sized FPM/worker pools, firewall and WAF rules, per-tenant Redis ACLs and database isolation.
  • Ship regeneration-proof configuration as code using persistent vhost include layers, so platform-generated configs never overwrite custom tuning.
  • Automate repeat toil with Bash tooling and cron/supervisor pipelines; maintain deployment workflows for Git-backed PHP, Node and Docker sites.
  • Write the team's incident runbooks and post-mortems, turning one-off fixes into repeatable, documented procedures.
LinuxNginxOpenLiteSpeedDockerPHP-FPMMySQLRedisCloudflareBashPM2

Implementation Engineer @ SouthTech Limited

2023 — 2025
  • Deployed and configured customer environments across hosting, email and DNS.
  • Ran server and site migrations with post-migration verification and rollback plans.
  • Built internal documentation and runbooks for recurring incidents.
LinuxWordPressDNScPanelBash
04.case studies

Real incidents — problem, diagnosis, durable fix.

problem

Sites across a shared VPS intermittently 502'd with DB-connection errors; mysqld was being OOM-killed repeatedly and stayed down until a manual restart.

diagnosis

The database was the victim, not the hog: one PHP-FPM pool with an oversized min_spare pinned 50+ idle workers and thrashed swap until the kernel killed the lowest-priority process. The stock unit's Restart=on-abort never restarts after a SIGKILL, so it never came back on its own.

fix

Right-sized the pool's pm.* values and graceful-reloaded FPM, then shipped a systemd drop-in with OOMScoreAdjust=-600, Restart=on-failure and a start-limit breaker — the DB now survives the spike and self-heals. Rolled the same drop-in across the fleet as a standard.

MySQLPHP-FPMsystemdCapacity
05.education

B.Sc. in Computer Science & Engineering

2010-2014

Ahsanullah University Of Science & Technology

Foundations in networking, operating systems, and databases that underpin my infrastructure work today.

06.contact

$ ./say-hello.sh

Got infrastructure to tame, or a DevOps role to fill? Let's talk.

The fastest way to reach me is email. I read everything and reply quickly.

Send a message