Bug: Excessive network egress / outbound bandwidth since v2.11.0

### 1. Default Sub-Minute Schedules & KV Store History Bloat (from Debug Logs)
The agent executes `docker-client-refresh` and `auto-heal` every 30 seconds (`*/30 * * * * *`). On every single tick, it updates a KV entry with an unpruned execution history containing dozens of past runs and receipts, generating large payloads:

Sep 19 22:18:00.060 DBG SQL query executed [19.02ms] query="UPDATE `kv` SET `value`='{\"jobId\":\"docker-client-refresh\",\"schedule\":\"*/30 * * * * *\",\"nextRun\":\"2026-09-19T22:18:30+02:00\",\"runs\":[{\"id\":\"271a30d9...\",\"status\":\"succeeded\"}, ... dozens of historical runs ...],\"receipts\":{...}}' WHERE key = 'jobs.df03bfe23f26e657fc7a2cbc5aad713658c1aa3b17137039b9d9eed4386b1d32'"
Sep 19 22:18:00.086 DBG SQL query executed [4.57ms] query="UPDATE `kv` SET `value`='{\"jobId\":\"auto-heal\",\"schedule\":\"*/30 * * * * *\",\"nextRun\":\"2026-09-19T22:18:30+02:00\",\"runs\":[{\"jobId\":\"auto-heal\",\"status\":\"succeeded\"}, ... dozens of historical runs ...]}' WHERE key = 'jobs.1a433b118e8afafd01123e2a0519d7775a7c6dabf6157a47c15a03190c420019'"
Sep 19 22:18:00.116 DBG Docker client refresh completed
Sep 19 22:18:00.122 DBG SQL query executed query="UPDATE `activities` SET `duration_ms`=104,`latest_message`='Auto Heal: succeeded',`status`='success' WHERE id = 'd6a022f5-9f0e-5ac2-b2eb-7f6e316fb602'"

### 2. GitOps Sync <-> Continuous Filesystem Watcher Loop
Every GitOps evaluation touches Compose files on disk, which immediately wakes the continuous filesystem watcher and triggers an immediate secondary sync cycle:

Sep 19 20:03:20.990 INF Using single file sync mode syncId=1db15a73 composePath=docker/uptimekuma_gcp_compose.yml
Sep 19 20:03:21.058 INF Using single file sync mode syncId=74d8be79 composePath=docker/netbird_gcp_01_compose.yml
Sep 19 20:03:21.114 INF Using single file sync mode syncId=b6fef13e composePath=docker/cftunnel_gcp_compose.yml
Sep 19 20:03:21.146 INF Updated project files projectName=uptime_kuma_gcp
Sep 19 20:03:21.749 INF Filesystem change detected, syncing projects
Sep 19 20:03:21.795 INF Project sync completed after filesystem change

### 3. Hardcoded 2-Minute Sync Workers
The agent repeatedly queries remote registry, S3, and repository endpoints on a fixed 120-second cycle:

Sep 19 20:04:00.371 INF Incoming request method=POST path=/api/container-registries/sync duration=8.27ms status=200
Sep 19 20:04:00.413 INF Incoming request method=POST path=/api/backups/s3/sync duration=639µs status=200
Sep 19 20:04:00.463 INF Incoming request method=POST path=/api/git-repositories/sync duration=954µs status=200
Sep 19 20:04:00.503 INF Incoming request method=PUT path=/api/environments/0/templates/variables duration=540µs status=200

### 4. 5-Second UI Polling Over WAN
When the Arcane web dashboard is open, polling occurs every 5 seconds, transmitting the bloated activities state over the network:

Sep 19 20:00:39.698 INF Incoming request method=GET path=/api/environments/0/activities duration=44.81ms status=200
Sep 19 20:00:44.694 INF Incoming request method=GET path=/api/environments/0/activities duration=45.68ms status=200
Sep 19 20:00:49.785 INF Incoming request method=GET path=/api/environments/0/activities duration=141.97ms status=200
Sep 19 20:00:50.012 INF Incoming request method=GET path=/api/environments/0/dashboard duration=373.66ms status=200

### 5. Measured Network Impact & Isolation
* Baseline outbound throughput on host interface (ens4): ~230 KB/s continuous (~19.8 GB egress per 24 hours).
* Network isolation: Bridge `br-7f2b430c3e8c` (`arcane_agent_default`) accounted for 3.67 GB of 3.80 GB total VM egress (~96.5%).
* Direct verification: Executing `docker stop arcane_agent` immediately drops host outbound egress from ~230 KB/s to 0–1 KB/s.

services:
  arcane-agent:
    image: ghcr.io/getarcaneapp/arcane-headless:latest
    container_name: arcane_agent
    restart: unless-stopped
    volumes:
      - /var/run/docker.sock:/var/run/docker.sock
      - arcane_agent_data:/app/data
      - arcane_agent_cert:/app/certs:ro
      - /docker-data/projects:/docker-data/projects
    ports:
      - 3553:3553
    extra_hosts:
      - "arcane.strasser.internal:ip"
    environment:
      - PUID=0
      - PGID=0
      - TZ=Europe/Vienna
      - SSL_CERT_FILE=/app/certs/rootCA.pem
      - AGENT_MODE=true
      - AGENT_TOKEN=arc_removed
      - MANAGER_API_URL=https://arcane.strasser.internal:3552
      - ENVIRONMENT=production
      - ALLOW_DOWNGRADE=false
      - LOG_LEVEL=info
      - LOG_JSON=false
      - ANALYTICS_DISABLED=false
      - PROJECTS_DIRECTORY=/docker-data/projects
      - ARCANE_BACKUP_VOLUME_NAME=arcane_backup

volumes:
  arcane_agent_data:
    name: arcane_agent_data
    driver: local
  arcane_agent_cert:
    name: arcane_agent_cert
    driver: local