blumeops/fly/start.sh
Erich Blume 6e37abda5d C1: deploy adelaide-baby-shower-app to ringtail k3s
Adds the Adelaide / Heidi / Addie baby shower app — a Django guest
splash, raffle picker, and prize-assignment console — on ringtail k3s.
Public landing at shower.eblu.me (via fly proxy), tailnet admin at
shower.ops.eblu.me. App source: forge.eblu.me/eblume/adelaide-baby-shower-app,
wheel-published to the Forgejo Packages PyPI index.

Manifests under argocd/manifests/shower/: NFS-backed PVC for /app/media,
local-path PVC for SQLite, ExternalSecret pulling DJANGO_SECRET_KEY from
1Password (item "Shower (blumeops)"), Tailscale ProxyGroup ingress.

Defense-in-depth for the public surface:
  - /admin/ blocked at the fly edge except /admin/login/ and /admin/logout/
  - shower_auth rate limit on the login path
  - new fail2ban filter+jail with a per-service shower-deny.conf
    (nginx-deny action generalized to accept nginx_deny_file)
  - django-axes (5 / 1h) keyed on (username, ip_address)

Plus: Caddy route on indri, Pulumi gandi CNAME, Grafana APM dashboard
mirroring docs-apm.json, runbook at how-to/operations/shower-app.md,
and a service-versions entry. X-Clacks-Overhead set on the new server
block — GNU Terry Pratchett.

Build: containers/shower/default.nix uses dockerTools to ship a
nixpkgs Python plus a startup wrapper that installs the wheel into
/app/data/.venv on first boot and execs gunicorn. Lets the wheel come
from forge PyPI without pinning hashes for every transitive dep.

Prerequisites tracked in the runbook (not yet executed):
  - NFS share sifaka:/volume1/shower (manual Synology step)
  - 1Password item "Shower (blumeops)" with secret-key field
  - container build via `mise run container-build-and-release shower`
  - Pulumi dns-up after merge
  - fly certs add shower.eblu.me

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
2026-05-11 08:14:12 -07:00

47 lines
1.6 KiB
Bash

#!/bin/sh
set -e
# Connect to tailnet first — nginx needs MagicDNS for upstream resolution.
# With bluegreen deploys, the old machine serves traffic until this one is
# fully ready. Fly.io runs Firecracker microVMs that support TUN devices
# natively — no need for --tun=userspace-networking.
tailscaled --statedir=/var/lib/tailscale --port=41641 &
sleep 2
tailscale up --authkey="${TS_AUTHKEY}" --hostname=flyio-proxy
until tailscale status > /dev/null 2>&1; do sleep 1; done
echo "Tailscale connected"
# Wait for MagicDNS to be ready — upstream blocks resolve DNS at config
# load, so nginx will fail to start if MagicDNS can't resolve yet.
echo "Waiting for MagicDNS..."
until nslookup forge.tail8d86e.ts.net 100.100.100.100 > /dev/null 2>&1; do
sleep 1
done
echo "MagicDNS ready"
# Ensure fail2ban per-service deny files exist before nginx starts
# (the geo directive's `include` fails if the file is missing).
touch /etc/nginx/forge-deny.conf
touch /etc/nginx/shower-deny.conf
# Start nginx — MagicDNS is available, upstreams resolved.
nginx -g "daemon off;" &
NGINX_PID=$!
echo "Nginx started"
# Start fail2ban for login brute-force protection.
# Non-fatal — nginx rate limiting is the primary defense; fail2ban is additive.
if fail2ban-server -b; then
echo "fail2ban started"
else
echo "WARNING: fail2ban failed to start (nginx rate limiting still active)"
fi
# Start Alloy for observability (logs → Loki, metrics → Prometheus)
alloy run /etc/alloy/config.alloy \
--server.http.listen-addr=127.0.0.1:12345 \
--storage.path=/tmp/alloy-data &
echo "Alloy started"
# Block on nginx — container exits if nginx stops
wait $NGINX_PID