~/about · whoami

Systems and buildings.

I'm Ahmet Dereli: 13 years in infrastructure, ten of them on a public cloud, from the hypervisor up. Today I keep many websites and cloud accounts patched and monitored, and build products of my own. On the other track I work as a building contractor: houses and villas in Antalya and Burdur.

Two tracks Software and building.

infra · since 2013

Servers and software

Ten years on a public cloud, from the hypervisor up; the last of them leading a team of 20+ engineers. Now many websites and cloud accounts, and products of my own.

Software

build · contractor

Houses and villas

Houses and villas as a building contractor. A single-storey mountain house in Bucak was built in 2023 and has its occupancy permit; a villa with a pool in Kemerağzı has its design done and its structure starting.

Construction

What I do The infrastructure side.

01 · cloud & platform

From servers to cloud

Hypervisors, VM images, storage and network automation for a public cloud. Parts of it shipped as products: a network gateway, an object-storage gateway, geo-DNS acceleration.

02 · running many sites

Many sites, one way of doing things

Dozens of Drupal, WordPress and CiviCRM sites on PaaS and VMs, kept consistent with ansible-pull, 50+ roles and a library of checks.

03 · security at scale

Security updates

When an advisory comes out: find which sites are really affected, patch them in a safe order and check the patch reached every one.

04 · cloud migrations

Cloud migrations

Multi-account cloud set up as code, central identity, audit trails with locked logs, and production moved onto new standards through reviewed plans.

05 · data & delivery

Data and delivery pipelines

ETL into a warehouse, and CI/CD with approval gates, drift checks and rollback, for apps and analytics alike.

06 · leading & selling

Team and sales

Led a team, hired and trained, sat in pre-sales meetings and gave a talk on high availability.

Path So far.

  1. now

    Building contractor

    Houses and villas. A mountain house in Bucak built in 2023, occupancy permit issued; a villa with a pool in Kemerağzı designed, structure starting.

  2. 2022 — now

    DevOps engineer, web platforms

    Many Drupal, WordPress and CiviCRM sites: patching, upgrades, CI/CD, mail and incident work. Alongside it, cloud landing zones, data pipelines and delivery for larger teams.

  3. 2019 — 2022

    Engineering lead, public cloud

    Led a team of 20+ engineers serving 1,000+ customer environments. Owned the network and storage products, joined pre-sales, hired and trained.

  4. 2015 — 2019

    Automation & DevOps engineer

    Ansible for the whole stack: hypervisors, images, networking, DNS, monitoring. Ran a small R&D group.

  5. 2013 — 2015

    Where it started

    Joined the same cloud company while still at university. My first servers and first outages were there.

  6. before

    The groundwork

    Network systems at a technical high school, network courses, and the first servers I set up, and broke, at home.

Talk & education In short.

talk · 2021

Understanding High Availability

Failover, ARP, virtual IPs with VRRP, and an HAProxy cluster as the worked example.

education

Computer technology and information systems

BSc in computer technology and information systems; graduate studies in business, cyber security and computer engineering.

Toolbox Tools I use.

automation

  • ansible
  • terraform
  • bash
  • python
  • go
  • make

platforms

  • linux
  • xen
  • docker
  • kubernetes
  • helm
  • zfs

cloud

  • multi-account iam
  • object storage
  • serverless
  • managed kubernetes
  • paas

delivery

  • ci pipelines
  • jenkins
  • argocd
  • composer

observability

  • prometheus
  • grafana
  • alertmanager
  • netdata
  • elk
  • zabbix

data

  • mysql / mariadb
  • postgresql
  • redis
  • mongodb
  • kafka
  • rabbitmq

web stack

  • nginx
  • openresty
  • haproxy
  • varnish
  • drupal
  • wordpress
  • civicrm
  • php

network

  • openvpn
  • ipsec / strongswan
  • wireguard
  • powerdns / bind

How I work A few principles.

  1. Read before you touch.

    Understand what the system does today before changing what it should do tomorrow.

  2. Fix the cause, not the symptom.

    If one server broke, I check whether the same problem is waiting on the others.

  3. One host, then everyone.

    A dry run, then a small group, then all of them. Even for a small change.

  4. Write it down.

    A fix that is not written down has to be found again next time.