vh/esh-pfi-infrastructure
1
0
Fork 0
Code Issues 7 Pull Requests Actions Packages Projects Releases Wiki Activity
Files
2dd459d2e5c9b81379cdb51a234ae413154d3fd8
esh-pfi-infrastructure/playbooks/homepage-regroup-irv-ml1.yaml
T
vh 39da1d4a97 feat(homepage): recategorise on "do I open this?", collapse the API groups 
The board mixed tools with endpoints. A vLLM seat whose href is a /docs page
sat in the same band as ComfyUI; the MQTT broker and the RustDesk relay, which
have no page at all, sat in Apps; and `Service Networking` was thirteen members
spanning three AdGuards, five Dockges, two Traefiks and four headless agents.

Every group is now one of two kinds and they never mix. TOOLS are expanded and
sit at the top of their tab. ENDPOINTS — an API, a broker, a background agent,
an href that is /docs or /ping or nothing — carry `initiallyCollapsed: true`
and sit at the bottom. Collapsed is not hidden: the eyebrow and its rule still
render, so the tab still says the thing exists and one click expands it.

A second rule fell out of the same pass and now shapes the group boundaries: a
group's members should all carry a widget or none should. A stat strip makes a
card ~50px taller, so one widget card in a row of plain ones opens a void under
the plain ones. That is why AdGuard and Traefik get their own groups rather
than sharing one with Dockge, and it is most of why the old Service Networking
band looked broken. AdGuard (ANA) was the last short card in its row and now
carries the same query/blocked/latency strip as its two siblings — one
infra-ops AdGuard login authenticates against all three instances, verified
against each; it lives in that stack's .env on the host and is vaulted.

The sixteen GPU-backed model seats were deliberately NOT relabelled.
`homepage.group` is read at container creation, so clearer names for
`AI - Inference` and friends would have cost a recreate on six vLLM seats, four
eval seats and four TTS engines — multi-minute model reloads on endpoints peers
reach through the gateway. Order plus `initiallyCollapsed` buys the same
separation for nothing, so those names stay as they are on purpose.

28 containers that ARE cheap to bounce were relabelled, across five hosts, via
rerunnable elway playbooks. Their label steps are gated on the old value still
being present, so a second run reports skipped rather than churning. Two verify
steps were wrong on first contact and are fixed with the reason recorded: the
traefik check raced its own recreate, and asserting a model seat is "running"
cannot answer "did I bounce it" when a seat may be legitimately stopped —
container age can, and now does.

The canonical stacks/ tree was synced to the deployed labels afterwards, so
intent and reality agree again on all fourteen tracked stacks.

Also documents the real nature of the post-recreate blank dashboard, which cost
~25 minutes here and an hour on 2026-08-19. `initialSettings":{}` in the served
HTML is the catch branch of the page's data loader, not a warm-up and not a
cache — and the error can vanish entirely, because the logger is assigned inside
the same try and the catch only logs if the logger exists. Ruled out by
measurement this time: all four API routes return 200 with correct content while
the page serves {}, and the previous known-good settings.yaml reproduces it
identically. The README now carries the one-command test and the next lead.

Before/after, all four tabs: http://10.100.10.50:8090/b/homepage-relayout/
2026-08-24 08:54:06 -07:00

116 lines
4.8 KiB
YAML
Raw Blame History

# Homepage recategorisation — irv-ml1 (10.100.79.3), 5 containers.
# Sibling of playbooks/homepage-regroup-ana-docker.yaml; rationale lives there.
#
# arbo, comfyui, waterland-studio -> AI - Studios (was AI - Image & Media)
# yt-voice-clipper -> AI - Studios (was AI - Audio Tools)
# dockge -> Compose Consoles
#
# `AI - Studios` is the "you open this and do work in it" group; the ASR and
# TTS API seats stay where they are and get collapsed by settings.yaml instead,
# which is what keeps the GPU seats out of this playbook entirely.
#
# ⚠ yt-voice-clipper carries its homepage labels in `docker-compose.override.yml`,
# not in `docker-compose.yml`, and its compose dir is a git checkout of the
# project — the override is the deploy-local layer, which is the right place
# for it.
#
# ⚠ This host is reached over the WireGuard tunnel. If the run cannot connect,
# check the tunnel before assuming the host is down.
#
# Run: scripts/elway infra-ops@10.100.79.3 --playbook playbooks/homepage-regroup-irv-ml1.yaml
steps:
- name: arbo -> AI - Studios
sudo: true
shell: >-
sed -i 's|homepage.group=AI - Image . Media$|homepage.group=AI - Studios|'
/opt/docker/compose/arbo/compose.yaml
when: grep -q 'homepage.group=AI - Image . Media$' /opt/docker/compose/arbo/compose.yaml
- name: comfyui -> AI - Studios
sudo: true
shell: >-
sed -i 's|homepage.group=AI - Image . Media$|homepage.group=AI - Studios|'
/opt/docker/compose/comfyui/compose.yaml
when: grep -q 'homepage.group=AI - Image . Media$' /opt/docker/compose/comfyui/compose.yaml
- name: waterland-studio -> AI - Studios
sudo: true
shell: >-
sed -i 's|homepage.group=AI - Image . Media$|homepage.group=AI - Studios|'
/opt/docker/compose/waterland-studio/compose.yaml
when: grep -q 'homepage.group=AI - Image . Media$' /opt/docker/compose/waterland-studio/compose.yaml
- name: yt-voice-clipper -> AI - Studios
sudo: true
shell: >-
sed -i 's|homepage.group=AI - Audio Tools$|homepage.group=AI - Studios|'
/opt/docker/compose/yt-voice-clipper/docker-compose.override.yml
when: grep -q 'homepage.group=AI - Audio Tools$' /opt/docker/compose/yt-voice-clipper/docker-compose.override.yml
- name: dockge -> Compose Consoles
sudo: true
shell: >-
sed -i 's|homepage.group=Service Networking$|homepage.group=Compose Consoles|'
/opt/docker/compose/dockge/compose.yaml
when: grep -q 'homepage.group=Service Networking$' /opt/docker/compose/dockge/compose.yaml
# ---- recreates ---------------------------------------------------------
- name: recreate arbo
sudo: true
shell: cd /opt/docker/compose/arbo && docker compose up -d engine
- name: recreate comfyui
sudo: true
shell: cd /opt/docker/compose/comfyui && docker compose up -d comfyui
- name: recreate waterland-studio
sudo: true
shell: cd /opt/docker/compose/waterland-studio && docker compose up -d waterland-studio
- name: recreate yt-voice-clipper
sudo: true
shell: cd /opt/docker/compose/yt-voice-clipper && docker compose up -d api
- name: recreate dockge
sudo: true
shell: cd /opt/docker/compose/dockge && docker compose up -d dockge
verify:
- name: every relabelled container now carries its new group
sudo: true
changed_when: "false"
shell: >-
docker inspect -f '{{.Name}} {{index .Config.Labels "homepage.group"}}'
arbo comfyui waterland-studio yt-voice-clipper-api-1 dockge
- name: no container is left in the retired Service Networking group
sudo: true
changed_when: "false"
shell: >-
test -z "$(docker ps -q --filter 'label=homepage.group=Service Networking')"
# Assert NOT-RECREATED, not RUNNING. The first version of this checked that
# all five seats were up and failed — because chatterbox-fast has been down
# since 2026-08-10 and speaches since earlier the same morning, both long
# before this playbook existed. "Is it running" is the wrong question: a seat
# can be legitimately stopped. The question this step is actually asking is
# "did I bounce a model seat to relabel a dashboard", and container age
# answers it directly.
- name: the TTS and ASR seats were NOT recreated by this run
sudo: true
changed_when: "false"
shell: >-
for c in kokoro dots-tts chatterbox-fast tts-gateway parakeet speaches; do
created=$(docker inspect -f '{{.Created}}' "$c" 2>/dev/null) || continue;
age=$(( $(date +%s) - $(date -d "$created" +%s) ));
if [ "$age" -lt 600 ]; then echo "$c was recreated ${age}s ago"; exit 1; fi;
done
- name: everything relabelled is running
sudo: true
changed_when: "false"
shell: >-
test "$(docker inspect -f '{{.State.Running}}' arbo comfyui
waterland-studio yt-voice-clipper-api-1 dockge | sort -u)" = "true"
Reference in New Issue View Git Blame Copy Permalink