The image has no shell (ko-build distroless; entrypoint /ko-app/anubis), and this host's quadlet generator joins a bare HealthCmd into one \x20-escaped argv element that podman then runs shell-form — exec of a missing shell exits 1 and the container sits unhealthy. A JSON array in HealthCmd passes through the generator verbatim and executes the binary directly.
62 lines
2.8 KiB
INI
62 lines
2.8 KiB
INI
# Quadlet unit: Anubis bot-challenge proxy (rootless, boot-persistent).
|
|
#
|
|
# Install path on dankclaw: ~/.config/containers/systemd/anubis.container
|
|
# Unit name: anubis.service (container name: anubis)
|
|
#
|
|
# Two-tier chain: the public Caddy listeners for the guarded vhosts proxy
|
|
# everything to this container (port 8923, internal only — no published port,
|
|
# so metrics on :9090 stay reachable only inside dankclaw_default); Anubis
|
|
# verifies the client (challenge/DENY per policy) and forwards accepted
|
|
# traffic back to Caddy's internal listener (TARGET below), where the normal
|
|
# per-vhost routing — including Authentik forward_auth — resumes. Host headers
|
|
# survive both hops, so vhost matching on the internal listener still works.
|
|
#
|
|
# Image pin: the image is pulled from ghcr.io/techarohq/anubis (source of
|
|
# truth), but the unit runs it under a localhost/-aliased tag: on this host's
|
|
# podman 5.7 quadlet path, registry-prefixed refs fail to resolve from local
|
|
# storage at container create time ("image not known") even when present; the
|
|
# localhost alias resolves (verified live 2026-08-28). v1.13.0 was
|
|
# source-verified before pinning (policy parser is JSON-only at this tag,
|
|
# flags are the config surface via flagenv: TARGET sets --target,
|
|
# POLICY_FNAME sets --policy-fname). Upgrading = pull new tag, re-tag the
|
|
# localhost alias, restart the unit; re-verify the policy schema first.
|
|
# (the active Image= key lives in the [Container] section below)
|
|
#
|
|
# No secrets here: Anubis holds no credentials (memory store; challenge
|
|
# cookies are client-side JWTs). A restart invalidates issued challenges —
|
|
# users just re-challenge once. See serverAppConfigs-docs/Infrastructure/Anubis.md.
|
|
|
|
[Unit]
|
|
Description=Anubis bot-challenge proxy container
|
|
After=network-online.target
|
|
Wants=network-online.target
|
|
|
|
[Container]
|
|
ContainerName=anubis
|
|
Image=localhost/techarohq/anubis:v1.13.0
|
|
Pull=never
|
|
Network=dankclaw_default
|
|
# flagenv maps env names to flag names (TARGET -> --target, POLICY_FNAME ->
|
|
# --policy-fname). TARGET points back at Caddy's internal listener; the
|
|
# container DNS name resolves on dankclaw_default.
|
|
Environment=TARGET=http://caddy:8016
|
|
Environment=POLICY_FNAME=/policy/policy.json
|
|
Volume=/home/dank/serverAppConfigs/dankclaw/anubis/policy.json:/policy/policy.json:ro
|
|
# The image ships no HEALTHCHECK (distroless ko-build: entrypoint is
|
|
# /ko-app/anubis, no shell in the image), so the probe is declared here.
|
|
# JSON exec form on purpose: this quadlet generator joins HealthCmd into a
|
|
# single \x20-escaped argv element and podman then runs it as shell-form —
|
|
# which needs a shell the distroless image does not have. A JSON array is
|
|
# passed through verbatim and executes /ko-app/anubis directly.
|
|
HealthCmd=["/ko-app/anubis","--healthcheck"]
|
|
HealthInterval=15s
|
|
HealthTimeout=5s
|
|
HealthRetries=3
|
|
|
|
[Service]
|
|
Restart=always
|
|
RestartSec=5
|
|
|
|
[Install]
|
|
WantedBy=default.target
|