MosswartOverlord/go-services/docker-compose.go.yml
Erik 27757636e4 feat(go-services): tracker WS servers (/ws/position + /ws/live) + robust shadow
Completes the Go tracker as a cutover-ready drop-in:
- wslive.go: browser broadcast hub with per-client subscribe filters (nil=all),
  request_dungeon_map replies, and command routing; auth = internal-trust or
  session cookie. The ingestor broadcasts every handled event to it.
- wsposition.go: plugin ingest server with X-Plugin-Secret/SHARED_SECRET auth
  (constant-time, fails closed, legacy fallback), register -> plugin_conns, and
  dispatch into the shared Ingestor. plugin registry for backend->plugin commands.
- main.go: statusRecorder.Unwrap() so coder/websocket can hijack through the
  logging middleware (WS handshakes failed without it); /ws/ bypasses HTTP auth.

Shadow consumer robustness (the harness was being evicted under the full
firehose): decouple socket read from processing — the read loop only copies raw
frames to a queue; a worker unmarshals + dispatches. JSON parsing in the read
loop was slowing it enough that Python's broadcast send errored and evicted us
(Read then blocked forever). Added a 25s read-deadline watchdog to self-heal.

Validated live: shadow /live online = 73 = production; telemetry sustained ~12/s,
0 drops, no eviction; and the shadow's /ws/live re-broadcast stream is IDENTICAL
to production's (TOTAL 2150=2150, every event type exact).

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-24 11:15:05 +02:00

128 lines
4.8 KiB
YAML

# Compose OVERRIDE that adds the Go services alongside the live Python stack.
# It only ADDS containers; it never modifies the tracked docker-compose.yml or
# any running Python service.
#
# Invoke from the repo root so the Compose project name resolves to
# "mosswartoverlord" (same as the live stack) and the new container joins the
# existing default network — letting it reach the `db` service by name:
#
# cd /home/erik/MosswartOverlord
# export BUILD_VERSION="$(date -u +%Y.%-m.%-d.%H%M)-$(git rev-parse --short HEAD)"
# docker compose -f docker-compose.yml -f go-services/docker-compose.go.yml \
# build dereth-tracker-go
# docker compose -f docker-compose.yml -f go-services/docker-compose.go.yml \
# up -d --no-deps dereth-tracker-go
#
# --no-deps keeps Compose from touching the already-running `db` (and anything
# else). The service is loopback-bound (127.0.0.1:8770); external reach is only
# ever via the host nginx `location /go/` block (added separately).
services:
dereth-tracker-go:
build:
context: ./go-services/tracker-go
args:
BUILD_VERSION: ${BUILD_VERSION:-dev}
image: dereth-tracker-go:local
container_name: dereth-tracker-go
ports:
- "127.0.0.1:8770:8770"
environment:
PORT: "8770"
# Read-only use of the same dereth TimescaleDB the Python tracker writes.
DATABASE_URL: "postgresql://postgres:${POSTGRES_PASSWORD}@db:5432/dereth"
INVENTORY_SERVICE_URL: "http://inventory-service:8000"
# Same signing key as the Python tracker so the same login cookie verifies
# on both during the parallel run.
SECRET_KEY: "${SECRET_KEY}"
LOG_LEVEL: "INFO"
depends_on:
- db
restart: unless-stopped
logging:
driver: "json-file"
options:
max-size: "10m"
max-file: "3"
# Go port of discord-rare-monitor. Consumes the SAME Python /ws/live firehose
# as the live Python bot. DRY-RUN by default (logs classifications, posts
# nothing) so it can't double-post. To parallel-test for real, set a TEST
# DISCORD_RARE_BOT_TOKEN + TEST channel IDs + DRY_RUN=0 here.
discord-rare-monitor-go:
build:
context: ./go-services/discord-go
args:
BUILD_VERSION: ${BUILD_VERSION:-dev}
container_name: discord-rare-monitor-go
environment:
DERETH_TRACKER_WS_URL: "ws://dereth-tracker:8765/ws/live"
MONITOR_CHARACTER: "Dunking Rares"
ICONS_DIR: "/icons"
LOG_LEVEL: "INFO"
# DISCORD_RARE_BOT_TOKEN: "" # set a TEST token to go live
# DRY_RUN: "0" # required (with a token) to actually post
# COMMON_RARE_CHANNEL_ID / GREAT_RARE_CHANNEL_ID / SAWATOLIFE_CHANNEL_ID /
# ACLOG_CHANNEL_ID: set TEST channels before going live
volumes:
- ./discord-rare-monitor/icons:/icons:ro
depends_on:
- dereth-tracker
restart: unless-stopped
logging:
driver: "json-file"
options:
max-size: "10m"
max-file: "3"
# ---- Phase 2: shadow ingest (fully isolated; production never touched) ----
# A SEPARATE TimescaleDB the Go tracker owns for shadow ingest. Isolated
# volume + loopback port; the production dereth DB is never written.
dereth-go-db:
image: timescale/timescaledb:2.19.3-pg14
container_name: dereth-go-db
ports:
- "127.0.0.1:5434:5432"
environment:
POSTGRES_DB: "dereth_go"
POSTGRES_USER: "postgres"
POSTGRES_PASSWORD: "${POSTGRES_PASSWORD}"
volumes:
- dereth-go-data:/var/lib/postgresql/data
restart: unless-stopped
logging:
driver: "json-file"
options:
max-size: "10m"
max-file: "3"
# Shadow tracker instance: same image, but OWNS dereth-go-db (read-write) and
# (once ingest lands) consumes the Python /ws/live firehose into it, so its
# ingest output can be compared against production without writing to it.
dereth-tracker-go-shadow:
image: dereth-tracker-go:local
container_name: dereth-tracker-go-shadow
ports:
- "127.0.0.1:8771:8771"
environment:
PORT: "8771"
DATABASE_URL: "postgresql://postgres:${POSTGRES_PASSWORD}@dereth-go-db:5432/dereth_go"
READ_ONLY: "false" # owns its DB; creates schema on boot
INVENTORY_SERVICE_URL: "http://inventory-service:8000"
SECRET_KEY: "${SECRET_KEY}"
SHARED_SECRET: "${SHARED_SECRET}" # /ws/position plugin auth (cutover-ready)
SHARED_SECRET_LEGACY: "${SHARED_SECRET_LEGACY:-}"
# Replay the Python /ws/live firehose into the ingest handlers (shadow).
SHADOW_INGEST_WS: "ws://dereth-tracker:8765/ws/live"
LOG_LEVEL: "INFO"
depends_on:
- dereth-go-db
restart: unless-stopped
logging:
driver: "json-file"
options:
max-size: "10m"
max-file: "3"
volumes:
dereth-go-data: