This page was last updated Wed, April 17th, 2019.
At home, I use a fully decked out new Mac Mini, with 64 GB of memory, with a NAS attached giving me approximately 20TB of storage with ~1ms latency. I mount this NAS on my Mac, giving me direct access.
At work, I use a variety of *nix based operating systems, primarily Ubuntu 16.04/18.04 + Debian Jessie/Stretch.
My lab is a mix of NAS (mentioned above), old PCs, a few tower servers. My rack mount gear was shipped to colocation due to energy, space, and noise issues. It is much more cost effective to ship it to colocate than keep it locally.
Hypervisor 01 (Ed)
32 GB DDR3 ECC / 3x4TB / 1x120G SSD / 1Gbit Unmetered Port / Proxmox (Serves as dev testbed) To be deprecated.
Hypervisor 02 (Borris)
64 GB DDR4 ECC / 2x10TB / 2x512G NVMe / 1Gbit Unmetered Port / Proxmox (Serves as current production workload) I want to scale this down, possibly remove from colocation, this hypervisor is overkill for my needs. To be deprecated.
Hypervisor 03 (Xelayan)
16 GB DDR4 ECC / 120 GB SSD / 2x2TB over NAS / 1 Gbit Unmetered Port / Debian (serves as Docker host, hosting Jira among other applications). Production.
Hypervisor 04 (Isaac)
16 GB DDR3 ECC / 120 GB SSD / 1x1TB over NAS / 1Gbit Unmetered Port / Debian (serves in production for headless API hosting) Production.
Hypervisor 05 (Kelly)
16 GB DDR3 ECC / 120 GB SSD / 1x1TB over NAS / 1Gbit Unmetered Port / Debian (serves in production for analytics, heatmap-style) Production.
Hypervisor 06 (Kylden)
8GB DDR3 ECC / 120 GB SSD / 1x1TB over NAS / 100Mbit Unmtered Port / Debian (serves as a test/staging bed for Isaac, Kelly) To be deprecated.
I use the following tools on a daily basis:
- Exchange Server & Active Directory
- PHPStorm / Webstorm
- Ghost (powers this blog)
I classify my list of programming languages by confidence level, this is not necessarily reflective of all the programming languages I know / have worked with.
- Typescript (preferred)
- C++ & C
- Ruby / Ruby On Rails
At home, I run two generic 24" ASUS HDMI monitors (1920x1080), side-by-side, with an additional 32" LG monitor above (1080p) that has my grafana + monitoring infrastructure open, so I can keep an eye on all my servers easily, in real time. Razer gaming keyboard + mouse.
At work, I use a Dell Precision laptop on a laptop stand, with two (I believe) 30" ASUS monitors, side-by-side on monitor arms. Generic keyboard (Logitech) + Wireless Mouse.
For all my global infrastructure, I use a mixture of Prometheus, Grafana to keep an eye on all production and development workloads. As I've started migrating some production workloads to DigitalOcean, Vultr, OVH I need to seek out how to better integrate the monitoring without using tons of bandwidth.
My dashboard consists of three dashboard sections (ironic, yes) - Production at the top, where I monitor real-time bandwidth in-out, proxy logs (nginx scraping) to monitor requests per second, VM count of each hypervisor, check disk stats (full/empty), and load averages, etc.
For staging, it's all gated off, however I monitor attempts to access (including common bruteforce attempts). I've begun programming scripts to automatically export data to send to IP abuse watchlists, and push them to my pfsense router for blocking at the edge. Now that I'm also using DigitalOcean, when I sync to production I'll need to push the rules to DO cloud firewall.
For development, I've left this gated to internal IPs only, and connect over wireguard vpn. I've whitelisted so only staging can access development, and my home IP address.
Goals for 2019
I'd love to get into making some videos on how I made this seemingly impossible dream I had just a few years ago of owning my own ISP + infrastructure a reality. I'll keep posting blog posts / tutorials on how I made this all possible in the mean time.
I also intend to do some live streaming this year, a misc collection of content like managing systems, migrating data in real time across DCs & countries, etc. Things are looking ahead quite well for this year, and I'm glad I can blog about it here.
I'm working on re-creating my personal infrastructure, to have entire oversight in real-time, alerts, and be on top of literally everything that's going on in the servers. By the end of 2019, I want all physical hypervisors deprecated in favour of being full container hosts. Then, I want all them hooked up to Prometheus and Grafana for entire oversight on the infrastructure.