Meetups/Infra/2026-02-06-Roundtable

Meeting Announcement

We'll be having a Digitial Infra Roundtable this Friday ~7:30pm to discuss near-term plans to improving various NB infra pain points. There will be pizza!

A loose agenda includes:

• Run through how current infra is structured / deployed
  • primary repo: github.com/noisebridge/infrastructure
  • m3, m5, m6, m7 are live as of today
  • Infra is (kind-of) configured via ansible.
  • Elan (paraphrased): We should enable people to hack on stuff and auto-deploy after a (trusted) review (or two)
    • Some discussion on deploying on merge vs deploying triggered via comments and merged upon successful completion. No decision yet.

Agenda check-in & priorities

• • let's talk about the pain points that brought us together. Wiki server hackability updates (& Cloudflare).

• agenda setting
• the intro

motivations:

• DNS
• Wiki
• Mailing lists
• Solicit ideas on how to improve our infra and get more people involved in contributing / curating cool NB services
  • Redundant data backups. The infra is not valuable, but unique data is.
  • Critical vs dev playgrounds.
  • K8s: Argo, Rancher,
  • Possible VM cloud service
  • Data servers?

Discussion

• What is noisegarden?
  • Bare metal kubernetes cluster
    • NoiseGarden

• What is Kubernetes / Docker / Containers?
  • Kubernetes: Orchestrates a bunch of docker containers on the available hardware.
  • Docker: Docker packages software into standardized units called containers that have everything the software needs to run including libraries, system tools, code, and runtime.
  • Containers

• Noisebridge remote servers.
  • Our infra is running in VMs remotely, not bare metal.

If I have a container with some application I'd like to run, how can I deploy it to the kubernetes cluster?

Topics to discuss:

• • what services do people want
  • internal domain names
  • external names, or NAT routing.
  • static ip ranges
  • visiblity of services on laptop
  • door fixing - visiblity of noisebridge
  • docker container for print room, hand drawn images into 3d printable, toy
  • ImageGen (live) brain dump of conversations
  • prod vs dev distinction -- local

Q

What services are lacking?

Naomi: "I would like to be able to set up an internal hostname w/ static IP address for myself"

Elan: Currently possible with a password

Naomi: Be able to set up my own subdomain @ noisebridge.net (or whatever) with (maybe) NAT-routing of a port into an internal machine (audited by peers yes)

• Static IP addresses?
• More visibility into infra
• Connecting on prem infra and remote infra.

Erikk:

• 2D -> 3D image service for the print room.
• Brain dump conversations
• Make some of the infra as non-essential more open - not require keys. Dev / sandbox area.

Sophia:

• transit alerts - when is next X showing up
• DIY doorbell cam / press a button upstairs to open it from upstairs (Perry: +1!)
• base model inference
• Flaschentashen web interface

Loren:

• wanting to deploy small services on NoiseGarden (inventory / limited todos)

Naomi:

• Home assistant on raspberri pi (internal)
• Streamlit dashboards w/ NAT-routed ports

Derek:

• locker management system w/ QR codes

Daniel:

• put hardware in local colo w/ donated hardware

How to modernize our noisebridge/infrastructure repo?

• Peter: What is the contract for hosted service?
  • Security (security review process?)
  • Externally accesible?
• Sophia: Does Docker solve this? Daniel: Unfortunately it is not a security layer
• Loren: We have limited network bandwidth to host externally visibility

Fine for internal-only things to be unsecured, but it's a nice excuse to learn how to use letsencrypt.

• can't do subdomains bc cookies are by-default shared

• Elan: Let's have 3 tiers re: trust:
  • "Noisebridge services" - all noisebridgers depend on them. (Wiki, etc.)
  • "Core services" - other services depond on them.
  • "community services" - do whatever you want.

• Erikk: Is there a list of our hardware specs?
  • Let's do this once a month ("office hours") (Naomi: "with pizza?")

• Erikk:
  • Need backups, data (esp wiki data) is critical. Can't imagine losing that.
  • VM rancher
  • Proxmox hypervisor, with terraform provisioning.

• John: "The only way you know you have backups is if you can use them in anger"
• Loren: That's how we can populate a staging/test env.
• Elan: Would love if we can save this and serve a copy from the nb local network, if the public wiki is down.
• Naomi: I use the api to make a backup 1/hour.
  • https://claude.ai/public/artifacts/76db8105-55f1-4c1c-b1af-33fc98047ff2

• Peter: autoschematic as a tool for testing deployments.