Auto-publish 2026-09-06_15:52 (#826)

https://projects.knownelement.com/issues/826
This commit is contained in:
2026-09-06 15:52:58 -05:00
parent 3e47443784
commit 0f3fb7e05e
8 changed files with 446 additions and 20 deletions
+19
View File
@@ -0,0 +1,19 @@
---
title: About
visibility: private
---
# 👥 About the performance organization
The KNEL Performance organization runs fleet performance optimization for
Turnsys: the baseline → audit → tweak → re-baseline loop, capacity
planning, and power-aware scheduling across two sites (main PFV datacenter
+ SiTES fabrication shop). It reports to the VP of Technical Operations
with a dotted line to the VP of Facilities.
- Work tracking: [Redmine #826](https://projects.knownelement.com/issues/826)
(program anchor ticket)
- Code: [KNELPerf](https://git.knownelement.com/KNEL/KNELPerf) ·
[KNELSiterSolar](https://git.knownelement.com/KNEL/KNELSiterSolar) ·
[this site's content](https://git.knownelement.com/KNEL/KNELPerfHub)
- License: AGPLv3 across the program
+25
View File
@@ -0,0 +1,25 @@
---
title: Fleet
visibility: private
---
# 🖥️ Fleet inventory
Hardware transparency: every node serving this program, verified by
live dmidecode survey (2026-09-06). Capacity planning and margin
accounting run against this table.
| node | platform | cores | RAM | storage character |
|---|---|---|---|---|
| pfv-tsys1 | OptiPlex 9020 (i7-4770) | 8 | 32G | 1x 1TB spinner — edge role |
| pfv-tsys3 | Precision 7510 (E3-1535M v5) | 8 | 32G | NVMe (PM961) — fast singleton |
| pfv-tsys4 | Precision T1700 (E3-1246 v3) | 8 | 16G | 6 spindles + 3 more USB toasters coming = spindle farm |
| pfv-tsys5 | Precision T7500 (E5620) | 8 | 96G | mixed; NFS server for -02 (NEVER casually rebooted) |
| pfv-tsys6 | PowerEdge R610 (2x E5530) | 16 | 128G | 1x 2TB spinner |
| pfv-tsys7 | PowerEdge R620 (2x E5-2630 v2) | 24 | 192G | 1x 2TB spinner |
| pfv-tsys8 (new) | Lenovo, NVMe, 12G | ? | 12G | fast NVMe, small — scheduler/quorum/edge |
| pfv-tsys2 (soon) | Precision 7510 clone + NVMe | 8 | 32G | NVMe |
Power telemetry per node: RAPL (tsys1/3/4/7/9) · iDRAC (tsys6/7) ·
metered PDU (tsys5) — methodology in the
[AMT report](/performance/reports/report-amt-power-telemetry).
+20
View File
@@ -0,0 +1,20 @@
---
title: ERCOT Risk Intelligence
visibility: private
---
## Day-to-day ISO risk intelligence (ercot-brief.sh)
Founder ruling 2026-09-06: Base Power must remain the retail provider (no
direct ERCOT interconnection). So the ISO's signals are RISK INTELLIGENCE,
not an export market:
- `ercot-brief.sh` produces a daily risk level from LZ_SOUTH RTM (now +
24h peak): GREEN (<$150/MWh) / WATCH (>=150) / ELEVATED (>=500) /
CRITICAL (>=2000, adders/EEA territory).
- Actions it emits: precharge battery (WATCH+), defer class-3 batch load
(ELEVATED+), rtw_counterfactual_flag (RTM > $85).
- On EEA-watch days expect Base to dispatch the fleet battery and expect
RTW-plan counterfactual spikes — history TSV accumulates for trend stats.
- Same source chain as the arbitrage tool (GRIDSTATUS_API_KEY /
ERCOT_API_KEY / override — key registration still pending, see #826/#827).
+67
View File
@@ -0,0 +1,67 @@
---
title: Solar Arbitrage
visibility: private
---
# ONCOR real-time arbitrage: sell to Base Power vs serve load
Status: v1 (2026-09-06, #826). Location: Pflugerville, Travis County, TX —
ONCOR wires, ERCOT South load zone (`LZ_SOUTH`). Tool:
`solar-analysis/oncor-arbitrage.sh`.
## The plan facts that shape everything
Base Power (verified 2026-09-06):
- **Import: $[redacted — LLC-private] flat** (contract through Oct 2028)
- **Export: $[redacted — LLC-private]/kWh flat** solar credit — Base does NOT pay real-time
wholesale for exports
So on Base Power, **both sides of your meter are flat rates**. ERCOT's
real-time market price (RTM, 15-min settlement point prices) does not change
what you pay or earn today — it is a *counterfactual* and a *planning
signal*:
## Decision rules (what the tool emits)
For each kWh of solar surplus, best action = max of:
| option | value | when it wins |
|---|---|---|
| serve own/deferrable load | $[redacted — LLC-private] avoided | almost always ([redacted] > [redacted]) |
| export to Base | $[redacted — LLC-private] flat | only when NO deferrable load exists |
| export on an RTW plan* | RTM $/kWh | only when RTM > $[redacted — LLC-private] (rare spikes) |
*counterfactual — would require switching export credit to a real-time-
wholesale retailer plan (e.g. Octopus-style RTW). Texas RTW export averages
~2¢/kWh with rare spikes >$5/kWh — as a steady diet it LOSES to Base's flat
[redacted]; it only wins if you can time exports into spikes (needs battery).
**Practical rules:**
1. **Every surplus kWh has a home: load first.** HFNOC batch compute, site-2
deferrable manufacturing, battery charging — all displace [redacted] power.
This is why the KNELPerf class-3 batch gate keys on solar surplus.
2. **Only truly surplus-after-load gets exported at [redacted].**
3. **RTM spike alerts** (`rtw_would_beat_self_consume: true`, RTM >
$85/MWh): under a future RTW plan you'd earn more exporting than
self-consuming. Track how often this happens; if frequent and clustered,
an RTW plan + battery beats Base's flat buyback for exports. Until then,
staying on Base and self-consuming is strictly better.
4. **Battery note**: Base's free battery — verify who controls dispatch
(Base dispatches it for grid services; if they own the dispatch, do not
count it in household arbitrage). Open question logged on #826.
## Price sources
Official API (api.ercot.com public API) and gridstatus.io both require a
free API key; ercot.com's dashboard JSON is behind an Imperva bot-wall
(tested 2026-09-06). Tool source precedence: `RTM_OVERRIDE`
`GRIDSTATUS_API_KEY``ERCOT_API_KEY` → cache file. Key registration is a
founder action ( NeedsInput on #826).
## KNELPerf integration
`oncor-arbitrage.sh` emits `start_deferrable_load` + `rtw_would_beat_self_consume`
as JSON — the batch launcher consumes this plus the HA solar-surplus sensor
once the plant is live. Docs: KNELPerf docs/ARCHITECTURE.md §7-8.
> **Docs live on Discourse — this repo is the executable source of truth.**
+15 -20
View File
@@ -5,30 +5,25 @@ visibility: private
# ⚡ KNEL Performance Hub
Auto-published 2026-09-06T15:28:59-05:00 from KNELPerf (#826). Repo:
https://git.knownelement.com/KNEL/KNELPerf
Live operational-performance intelligence for the Turnsys PFV fleet —
published openly, updated automatically. Ultra-transparent by
default; [revenue and expense detail stays LLC-private]
(/performance/transparency).
## ⚡ ERCOT risk (LZ_SOUTH)
- **What we do**: [the program](/performance/program) — a continuous
baseline → audit → tweak → re-baseline loop across a 9-node
Proxmox fleet, scheduled around solar surplus and grid risk.
- **The metal**: [fleet inventory](/performance/fleet).
- **Deep dives**: [reports](/performance/reports).
- **Methodology**: [grid-risk intel](/performance/intel/ercot-risk) ·
[solar arbitrage](/performance/intel/solar-arbitrage).
- **Right now**: [status dashboard](/performance/status).
## ⚡ ERCOT risk right now (LZ_SOUTH)
```json
{
"ok": false,
"error": "no price source (set GRIDSTATUS_API_KEY or ERCOT_API_KEY; override with RTM_OVERRIDE)"
}
{"error":"no source yet (API key pending)"}
{"error":"price feed pending API key registration"}
```
## 🌤️ Solar arbitrage verdict (latest run)
```json
{
"ts": "2026-09-06T20:28:59Z",
"ok": false,
"error": "no price source available (set ERCOT_API_KEY or GRIDSTATUS_API_KEY; see docs/oncor-realtime-arbitrage.md)"
}
{"error":"source unavailable"}
```
## 📋 Baseline index
| host | latest bundle |
|---|---|
| (none yet — first fleet baseline wave pending) | |
+253
View File
@@ -0,0 +1,253 @@
---
title: The Program
visibility: private
---
# 🏗️ KNELPerf v2 — architecture for dynamic, intelligent performance optimization
> Status: DRAFT v1 (2026-09-06,
> [#826](https://projects.knownelement.com/issues/826)). Founder directive:
> end-to-end baseline → audit/benchmark → tweak → re-baseline loop,
> two-week program.
## 📑 Contents
1. [The loop](#1-the-loop) · 2. [Scheduling architecture](#2-scheduling-architecture-ultix-breakout-redesign) ·
3. [Proxmox resource groups & VM balancing](#3-proxmox-resource-groups--vm-balancing) ·
4. [k8s scheduling & Slurm](#4-where-k8s-scheduling-fits-slurm) ·
5. [Telemetry spine](#5-telemetry-spine-beszel--power) ·
6. [Workload priority classes](#6-workload-priority-classes-founder-set-order) ·
7. [Solar + Home Assistant](#7-solar--home-assistant-integration) ·
8. [Two-site power economics](#8-two-site-power-economics-site-2-printer-farm--laser--cnc) ·
9. [Hardware plan hooks](#9-hardware-plan-hooks) · 10. [Organization](#10-organization)
## 1. The loop
```mermaid
flowchart LR
B[📋 BASELINE<br/>cpu-bench.tsv + bench-run.tsv<br/>+ PSI snapshot per host] --> A[🔍 AUDIT<br/>tolerance diff vs bundle<br/>beszel + PSI context]
A -->|drift found| T[🔧 TWEAK<br/>tweaks/&lt;id&gt;.sh — idempotent,<br/>GLPI CR-gated on prod]
T --> R[📈 RE-BASELINE<br/>capture + compare<br/>evidence to Redmine]
R --> A
A -->|clean| B
```
- **baseline**: `loop/perf-loop.sh baseline <host>` stores a bundle
(cpu-bench.tsv + bench-run.tsv + PSI snapshot) under `data/baselines/<host>/<ts>/`.
Storage/network legs reuse `bench/bench-run.sh` (fio + iperf3); the CPU leg
(`loop/cpu-bench.sh`) is new — the repo previously had zero CPU coverage.
- **audit**: `loop/perf-loop.sh audit <host>` re-runs the read-only probes and
diffs against the latest bundle via `loop/baseline-diff.sh` (tolerance-based,
catches governor flips exactly). This is what the scheduled loop runs.
- **tweak**: every mutation is a file under `tweaks/<id>.sh` (idempotent,
rollback helper included). Production tweaks ride a GLPI CR; the loop
refuses `tweak` on prod hosts without `.crush/active-cr`.
- **re-baseline**: same capture, then `compare` — the delta table is the
evidence artifact that lands on the Redmine ticket.
Cadence: nightly audit during the 22:0007:00 batch window (self-measuring:
the audit itself is low-cost); re-baseline only after a tweak or hardware
change. First full-fleet baseline wave is this week's deliverable.
## 2. 🗓️ Scheduling architecture (Ultix breakout redesign)
See `scheduler/README.md` for the port map and the lessons-learned list. Core
redesign decisions:
1. **Single source of truth per host**`/etc/knelperf/daynight.conf`; no
value ever lives in a timer file, unit default, or script constant.
2. **Reboot lands in day mode** — invariant carried from Ultix.
3. **Fleet-generic**: slice specs are data (`slice:weight:cpus:memhigh`), so
PVE hosts tune `qemu.slice` children and host services with the same engine
the workstation uses.
4. **Deployed by AWX (KNELIAC), not ad-hoc ssh** — scripts here are the
payload; job templates own the rollout waves.
## 3. 🖥️ Proxmox resource groups & VM balancing
Current fleet (dmidecode-verified 2026-09-06):
| node | platform | cores | RAM | storage character |
|---|---|---|---|---|
| pfv-tsys1 | OptiPlex 9020 (i7-4770) | 8 | 32G | 1x 1TB spinner — edge role |
| pfv-tsys3 | Precision 7510 (E3-1535M v5) | 8 | 32G | NVMe (PM961) — fast singleton |
| pfv-tsys4 | Precision T1700 (E3-1246 v3) | 8 | 16G | 6 spindles + 3 more USB toasters coming = spindle farm |
| pfv-tsys5 | Precision T7500 (E5620) | 8 | 96G | mixed; NFS server for -02 (NEVER casually rebooted) |
| pfv-tsys6 | PowerEdge R610 (2x E5530) | 16 | 128G | 1x 2TB spinner |
| pfv-tsys7 | PowerEdge R620 (2x E5-2630 v2) | 24 | 192G | 1x 2TB spinner |
| pfv-tsys8 (new) | Lenovo, NVMe, 12G | ? | 12G | fast NVMe, small — scheduler/quorum/edge |
| pfv-tsys2 (soon) | Precision 7510 clone + NVMe | 8 | 32G | NVMe |
Directions:
- **Proxmox resource groups / pools**: map pools to tenant classes (see §6),
so scheduler policy attaches to the pool, not per-VM.
- **Balancing across hosts AND spindles**: nodes are standalone (no
corosync/HA), so "balancing" = placement policy at create/migrate time, not
live DRRS. Initial heuristic: CPU-bound guests → tsys7 (24c) then tsys6;
RAM-hungry → tsys7/6/5; IO-latency-sensitive → NVMe nodes (tsys3, tsys8,
tsys2); sequential-throughput batch (HFNOC ETL) → tsys4 spindle farm with
one-job-per-spindle placement (fio baselines per spindle will quantify).
The audit loop's beszel/PSI data feeds a placement scorecard.
- **tsys6 vs tsys7 split**: RECOMMEND tsys7 = the big k8s worker (24c/192G —
largest), tsys6 = non-k8s VM farm (RackRental ContainerLabOLTP etc.).
Rationale: tsys7's Ivy Bridge has RAPL + iDRAC7 power telemetry (§5) and
the core count to soak k8s bins; tsys6's older CPUs are fine for
steady-state VMs but it lacks DCMI telemetry on old firmware. Decision
needs founder ruling — will be proposed on #826 with a capacity table once
tsys8/2 land.
## 4. ☸️ Where k8s scheduling fits; Slurm?
- **k8s (pfv-k8s, k3s)** owns long-running services + tenant classes
(ADR-0001 buckets). Resource requests/limits ARE the scheduling policy for
class 12 workloads; node labels/taints pin tenant classes to node pools.
- **HFNOC batch (class 3)** = interruptible, deadline-loose, large — this is
HPC-batch shaped. Options: (a) Kubernetes Jobs on spot-ish semantics
(preemption via priority classes — fits "start and stop easily"); (b) Slurm
on a dedicated partition across idle capacity. Recommendation: start with
k8s Jobs + PriorityClasses (no new control plane, flux-managed), and only
introduce Slurm if multi-node tightly-coupled MPI-style jobs appear. Slurm
co-located with k8s on shared nodes via cgroup slices is possible
(knelperf-batch.slice) but doubles scheduling systems for little gain at
this fleet size.
- The knelperf daynight engine gives the batch window (22:0007:00) its
expanded resource envelope on each participating node.
## 5. 📡 Telemetry spine: beszel + power
- **Beszel has 100% VM+physical coverage** → it is the always-on signal for
the audit phase (CPU/mem/disk/net per host). Plan: beszel agent metrics
feed the placement scorecard; PSI textfile collector
(`scheduler/collectors/`) adds pressure (queueing) semantics beszel lacks.
- **Power telemetry** (see docs/report-amt-power-telemetry.md): per-node watts
come from RAPL (tsys1/3/4/7) + iDRAC (tsys6/7) + metered UPS legs; tsys5
needs a metered PDU. UPS re-cabling plan (switches+tsys6/7 on the
un-metered UPS, everything else on the metered one) is compatible: the
nodes losing wall-meter visibility are exactly the RAPL/iDRAC nodes.
- **Load-side telemetry = per-circuit monitoring (#623, Emporia Vue 3
pending founder pick)**: per-circuit watts via the HA LOCAL (HACS)
integration → knelperf textfile exporter → the class-3 batch gate gets a
real deferrable-load menu (server-room subpanel vs HVAC vs site-2
equipment). Dedicated CTs on the server-room subpanel + the tsys6/7 UPS
legs keep fleet draw separable from house load — needed exactly on EEA
days, so local (non-cloud) telemetry matters.
## 6. 🏆 Workload priority classes (founder-set order)
| class | tenants | profile | scheduling |
|---|---|---|---|
| 1 | RackRental, Side Door Group, Starting Line (paid reservations) | OLTP, latency-sensitive, high margin | guaranteed reservations (k8s Guaranteed QoS / PVE cpuunits high); KillBill integration gives reservation calendar → capacity pre-reservation; promotions advised when solar surplus + spare capacity projected (§7) |
| 2 | Rogue Technologies (staff engagements), Suborbital Systems (R&D cost center) | bursty, business hours | Burstable QoS; preemptable by class 1 |
| 3 | HFNOC batch | GIS ETL, weather, time-series OLAP, ERP MRP/capacity-planning | night window (22:0007:00) or solar-surplus hours; checkpointable, preemptable by 12 |
## 7. ☀️ Solar + Home Assistant integration
- SITER-Solar (4 kW baseline design, PVWatts) → once the plant is live, Home
Assistant exposes real-time production. The loop consumes: surplus watts =
green light to start class-3 batches (excess production is the cheapest
compute on earth); forecast + reservation calendar (KillBill) → promotion
timing advice ("run the Side Door promo the week of X: projected surplus +
idle capacity Y").
- Integration path: HA REST/websocket (creds pattern exists:
`~/.creds/homeassistant.env`, tools/ha-ws-call.py in KNELBMS) → a small
exporter writing `knelperf_solar_surplus_watts` into the textfile
collector's directory; the batch launcher gates on it.
- SITER-Solar math review (2026-09-06): found constant drift
([redacted — LLC-private]/23952 vs plan's [redacted — LLC-private]/22614), a stale Dockerfile entrypoint
(deprecated python script), header/ROI inconsistencies, and fixed-cost
scenario assumptions — fixes landing in that repo.
## 8. 🏭 Two-site power economics (site 2: printer farm / laser / CNC)
The fleet has a second site with 3D-printer farm, laser cutter, and CNC — on
MUCH more expensive grid power than the main site. That asymmetry drives
scheduling policy as much as raw performance:
- **Marginal-price model**: every site has an effective $/kWh curve. Main
site = cheap grid + (soon) solar with export at $[redacted — LLC-private]and self-use at
$[redacted — LLC-private] (SITER-Solar contract). Site 2 = expensive grid. The scheduler's
objective becomes: run flexible load (compute AND manufacturing) where the
marginal $/kWh is lowest at that hour.
- **Solar surplus is main-site-first**: excess production self-consumed at
the main site displaces the CHEAPEST power — class-3 batch compute and any
deferrable main-site load should chase the surplus curve.
- **Site-2 manufacturing scheduling**: printer/laser/CNC jobs are batchy and
often deferrable (prints especially — hours-long, unattended). Candidate
policy: schedule site-2 energy-heavy jobs into site-2 off-peak windows
(if its tariff has TOU) or batch them for periods when main-site solar
surplus can't absorb more compute. Home Assistant + metered UPS legs give
the measured baseline; site 2 needs its own metering (smart plug / CT per
machine) before scheduling — queued as a data dependency.
- **Never shift OLTP**: class-1 paid workloads stay on their reserved
capacity regardless of power price — revenue and latency beat energy cost.
- KillBill reservation calendar × solar forecast × site-2 tariff = the
promotion/job-timing advisor output: "cheapest week to run X".
## 9. 🔩 Hardware plan hooks
- Purchases (RAM/CPU) deferred until after 2026-10-11; until then the wins
are placement + spindles (tsys4's 3 USB toasters join the spindle farm:
baseline each with fio, then one-batch-job-per-spindle).
- tsys8 (Lenovo NVMe 12G) → scheduler/edge roles + fast NVMe pool; inventory
row pending (PFVSystemInventory.csv not found in repo — founder to drop in).
- tsys2 (Precision 7510 clone + NVMe) → second fast singleton; NVMe pair
with tsys3 for latency-sensitive class-1 storage.
## 10. 👥 Organization
Performance management org reports to VPTechOps, dotted line VPFacilities —
identities per the TSGCOO pattern (Cloudron + Bitwarden + Redmine/Gitea/GLPI
groups). All work filed under #826; CRs for any prod change.
> **Docs live on Discourse — this repo is the executable source of truth.** Perf topic: https://community.turnsys.com/t/328
## 11. 📊 Capacity planning & margin tracking in Dolibarr
ERP landscape: 19 Dolibarr instances exist (one per brand);
**erp.knownelement.com** is the mothership (v24.0.0, REST API live, perf-org
API identity `vpperf` — creds `~/.creds/dolibarr-perf.env`). KillBill stays
the subscription billing engine; Dolibarr is the financial brain where
capacity, cost, and margin meet.
### Capacity ledger (supply)
- Fleet capacity is derived from ground truth we already collect: GLPI CMDB
+ dmidecode survey + baseline bundles → per-node vCPU/RAM/NVMe/spindle
inventory (ARCHITECTURE §3 table is the seed).
- Represented in Dolibarr as **virtual stock**: one product per capacity SKU
per node class (e.g. `cap-oltp-4c8g-nvme`, `cap-batch-spindle`,
`cap-k8s-worker-core`). Stock quantity = bookable units; GLPI drift report
reconciles it (the nightly reconcile already exists for CIs — extend to
capacity products).
### Reservations (demand)
- Paid reservations (class 1) = Dolibarr customer contracts/order lines that
consume capacity stock; KillBill invoices the money, Dolibarr holds the
commitment. Utilization % = booked/available per SKU per week.
- The promotion advisor consumes exactly this: KillBill reservation calendar
x projected solar surplus x utilization headroom (§7).
### Margin tracking (profitability)
- **Revenue** per tenant: KillBill (exports/journal) → Dolibarr customer
invoices/analytic accounts.
- **Direct costs** per tenant class: power (both sites — utility bills #827;
solar self-consumption at contract rate), hardware amortization (post
2026-10-11 purchases), WAN/tailnet overheads.
- Dolibarr **analytic accounting** (accounts + analytic distribution tables
are already present) allocates costs to tenant classes → margin per class
per month. PRIVATE dimension: margin numbers are revenue/expense detail —
LLC-members-only, never published to the public hub (transparency policy).
### Integration plumbing
- `loop/` baselines + beszel utilization → utilization updater script →
Dolibarr API (product stock / extrafields) — read/write via vpperf key.
- Hub publishes ONLY utilization % and capacity counts (public-safe);
margins stay inside the ERP.
API gotchas learned (2026-09-06): table prefix is `dolibarr_` (not `llx_`);
API auth needs `DOLAPIKEY` header; a cloned user needs explicit
`dolibarr_user_rights` rows copied (admin flag alone does not grant API
module rights).
+25
View File
@@ -0,0 +1,25 @@
---
title: Status
visibility: private
---
# 📈 Status dashboard
Auto-published 2026-09-06T15:52:58-05:00 from KNELPerf. Source:
https://git.knownelement.com/KNEL/KNELPerf
## 🌤️ Solar arbitrage verdict (latest run)
```json
{
"ts": "2026-09-06T20:52:58Z",
"ok": false,
"error": "no price source available (set ERCOT_API_KEY or GRIDSTATUS_API_KEY; see docs/oncor-realtime-arbitrage.md)"
}
{"error":"source unavailable"}
```
## 📋 Baseline index
| host | latest bundle |
|---|---|
| (none yet — first fleet baseline wave pending) | |
+22
View File
@@ -0,0 +1,22 @@
---
title: Transparency
visibility: private
---
# 🔓 Transparency policy
We run an ultra-open and transparent organization. This site and its
[content repo](https://git.knownelement.com/KNEL/KNELPerfHub) are **public
read-only** — performance data, benchmarks, risk intelligence,
architecture and decision reports are all in the open.
**The one carve-out**: revenue and expense detail is private to LLC
members and is never published here (system costs, billing history,
contract rates, per-tenant margins). This line is enforced mechanically:
the publishing script redacts financial literals from every public page
at render time — see
[`grav-publish.sh`](https://git.knownelement.com/KNEL/KNELPerf/src/branch/main/tools/grav-publish.sh).
Publishing is exclusively git-based: content lands via pull from the
public repo; the site runs no foreign tooling. The Grav admin exists only
to prevent first-run setup mode; content never flows through it.