bob
8b0eb0db78
JobQueue now skips bands swept within MinResweepHours (config, default 6h) instead of re-scraping the whole catalogue continuously — the dominant cost on the metered residential proxy. Roughly linear savings with no data loss (full pagination retained); 0 disables it. Worker logs the real compressed transferSize per job (what the proxy bills) rather than the ~6.5x-larger decompressed length, so spend is visible. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
56 lines
2.5 KiB
YAML
56 lines
2.5 KiB
YAML
# One-command startup for the cs.money scraper control plane + worker.
|
|
# Postgres is external (runs independently on the host); the C2 connects to it via
|
|
# host.docker.internal and auto-applies EF migrations on boot.
|
|
#
|
|
# docker compose up --build
|
|
#
|
|
# Scale workers (drop the worker `ports:` first — noVNC can't share one host port):
|
|
# docker compose up --build --scale worker=10
|
|
# Each worker mints its own IPRoyal sticky session at startup, so every replica gets a
|
|
# distinct residential exit IP. Set IPROYAL_USERNAME / IPROYAL_PASSWORD (e.g. in a .env
|
|
# file next to this compose file) to turn the proxy on.
|
|
services:
|
|
c2:
|
|
build:
|
|
context: .
|
|
dockerfile: BlueLaminate/BlueLaminate.C2/Dockerfile
|
|
environment:
|
|
# Point at the host's Postgres. Override the whole string for auth/host changes.
|
|
ConnectionStrings__SkinTracker: ${SKINTRACKER_CONN:-Host=host.docker.internal;Port=5432;Database=skintracker;Username=postgres}
|
|
WorkerToken: ${WORKER_TOKEN:-dev-worker-token}
|
|
MaxPagesPerJob: ${MAX_PAGES_PER_JOB:-60}
|
|
# Re-sweep floor (hours): skip bands swept more recently than this. The big lever
|
|
# for metered-proxy bandwidth — fewer redundant re-pulls. 0 = continuous re-sweep.
|
|
MinResweepHours: ${MIN_RESWEEP_HOURS:-6}
|
|
ports:
|
|
- "5080:5080"
|
|
extra_hosts:
|
|
# Lets the container resolve the host's Postgres on Linux too (no-op on Desktop).
|
|
- "host.docker.internal:host-gateway"
|
|
restart: unless-stopped
|
|
|
|
worker:
|
|
build:
|
|
context: .
|
|
dockerfile: worker/Dockerfile
|
|
environment:
|
|
C2_URL: http://c2:5080
|
|
WORKER_TOKEN: ${WORKER_TOKEN:-dev-worker-token}
|
|
# IPRoyal residential proxy: each replica self-assigns a unique sticky session
|
|
# (= unique exit IP). Auth is injected by an in-process forwarder, so no sidecar.
|
|
IPROYAL_USERNAME: ${IPROYAL_USERNAME:-}
|
|
IPROYAL_PASSWORD: ${IPROYAL_PASSWORD:-}
|
|
IPROYAL_COUNTRY: ${IPROYAL_COUNTRY:-us}
|
|
IPROYAL_LIFETIME_MIN: ${IPROYAL_LIFETIME_MIN:-60}
|
|
PROXY: ${PROXY:-} # auth-free host:port fallback (used only when IPRoyal creds are unset)
|
|
SOLVE_SECONDS: ${SOLVE_SECONDS:-45}
|
|
LOAD_IMAGES: ${LOAD_IMAGES:-} # set to 1 to re-enable images (debugging)
|
|
depends_on:
|
|
- c2
|
|
ports:
|
|
# Ephemeral host port so replicas don't collide under --scale. Find a worker's
|
|
# noVNC with `docker compose port worker 6080` (or `docker ps`), then open
|
|
# http://localhost:<mapped>/vnc.html to watch / solve a challenge.
|
|
- "6080"
|
|
restart: unless-stopped
|