docs(runbook): tg-digest VDS deploy runbook + stack artifacts (task:1311)

Ранбук первого деплоя паттерна sched+worker (модель для yt-digest): стек tg-digest (воркер internal, mem_limit 256m), runtime-регистрация задачи POST /tasks, env-контракт, smoke/verify/rollback, gotchas (сессия Telethon, TZ UTC, timeoutMs). Артефакты: compose source-of-truth, Dockerfile + http_worker.py (черновики для коммита в victor/tg-digest). task:1311 разблокирован (1305 done) -> ready.
This commit is contained in:
2026-08-28 23:33:47 +03:00
parent 73f2a74644
commit 9da36ddb27
5 changed files with 305 additions and 0 deletions

View File

@@ -0,0 +1,52 @@
# tg-digest — HTTP-воркер дневного дайджеста телеграм-каналов.
# sched cron 0 5 * * * (tz UTC = 08:00 MSK), runner http, simple mode.
# Internal: сеть proxy, БЕЗ traefik-labels — наружу не публикуется (см. ранбук tg-digest-vds-deploy-runbook).
# Deploy: Portainer-managed (см. portainer-stack-management-vds). Env через Portainer (не env_file).
# Image: registry.kzntsv.site/tg-digest-worker:<tag> (python:3.13-alpine, telethon+requests, из main victor/tg-digest).
services:
tg-digest:
image: registry.kzntsv.site/tg-digest-worker:<tag>
container_name: tg-digest
restart: unless-stopped
mem_limit: 256m
networks:
- proxy
environment:
# MTProto (оператор; pass telegram/api-id + api-hash)
TG_API_ID: ${TG_API_ID}
TG_API_HASH: ${TG_API_HASH}
TG_SESSION: /data/session.session
TG_PHONE: ${TG_PHONE}
TG_CACHE: /data/tg-cache
TG_WINDOW_HOURS: ${TG_WINDOW_HOURS:-24}
# Доставка (pass telegram/full-env)
TELEGRAM_BOT_TOKEN: ${TELEGRAM_BOT_TOKEN}
TELEGRAM_CHAT_ID: ${TELEGRAM_CHAT_ID}
# Стадия-2 LLM (deepseek; ключ — у оператора)
LLM_API_KEY: ${LLM_API_KEY}
LLM_BASE_URL: ${LLM_BASE_URL:-https://api.deepseek.com}
LLM_MODEL: ${LLM_MODEL:-deepseek-chat}
# Ингест raw в mappa (pass mappa/full-env)
MAPPA_URL: ${MAPPA_URL:-https://mappa.vds.kzntsv.site}
MAPPA_API_TOKEN: ${MAPPA_API_TOKEN}
MAPPA_PROJECT: ${MAPPA_PROJECT:-tg-digest}
# Auth sched → воркер (pass sched/tg-digest-api-key)
WORKER_API_KEY: ${WORKER_API_KEY}
# 0 — dry-run smoke без отправки в TG
TG_SEND: ${TG_SEND:-1}
volumes:
- tg-digest-data:/data
healthcheck:
test: ["CMD", "python", "-c", "import urllib.request;urllib.request.urlopen('http://127.0.0.1:8080/healthz')"]
interval: 30s
timeout: 5s
retries: 3
start_period: 15s
volumes:
tg-digest-data:
networks:
proxy:
external: true

View File

@@ -0,0 +1,21 @@
# tg-digest worker — python (telethon + requests) + HTTP-обёртка для sched http-runner.
# Файл коммитится в КОРЕНЬ репо victor/tg-digest (рядом http_worker.py).
# Build (из чекаута victor/tg-digest, main с d3aefb2):
# docker build -t registry.kzntsv.site/tg-digest-worker:<tag> .
# Env-контракт — см. ранбук tg-digest-vds-deploy-runbook (все секреты env, pass в контейнере нет).
FROM python:3.13-alpine
WORKDIR /app
COPY requirements.txt ./
RUN pip install --no-cache-dir -r requirements.txt
COPY src/ src/
COPY http_worker.py ./
RUN mkdir -p /data
ENV PYTHONUNBUFFERED=1
EXPOSE 8080
VOLUME ["/data"]
CMD ["python", "http_worker.py"]

View File

@@ -0,0 +1,63 @@
#!/usr/bin/env python3
"""tg-digest HTTP-воркер для sched (http runner, simple mode).
GET /healthz -> 200 {"ok": true} (healthcheck стека)
POST /run -> проверка x-sched-api-key (WORKER_API_KEY); запуск
`python -m src.worker`; 200 {ok, runId, summary} | 5xx {error}.
Simple mode: sched шлёт POST и ждёт ответ до config.timeoutMs; любой 2xx = succeeded.
runId приходит в заголовке x-sched-run-id (пишется в ответ, в лог).
Заголовки x-sched-api-key от sched — auth per-task (task config.auth.apiKey).
"""
import json
import os
import subprocess
import sys
from http.server import BaseHTTPRequestHandler, ThreadingHTTPServer
WORKER_API_KEY = os.environ.get("WORKER_API_KEY", "")
WORKER_CMD = [sys.executable, "-m", "src.worker"]
WORKER_TIMEOUT = float(os.environ.get("WORKER_TIMEOUT_S") or 3600)
PORT = int(os.environ.get("PORT") or 8080)
class Handler(BaseHTTPRequestHandler):
def log_message(self, *args): # тишина в stdout (логи — по runId)
pass
def _send(self, code, obj):
body = json.dumps(obj, ensure_ascii=False).encode("utf-8")
self.send_response(code)
self.send_header("Content-Type", "application/json; charset=utf-8")
self.send_header("Content-Length", str(len(body)))
self.end_headers()
self.wfile.write(body)
def do_GET(self):
if self.path == "/healthz":
self._send(200, {"ok": True})
else:
self._send(404, {"error": "not found"})
def do_POST(self):
if self.path != "/run":
return self._send(404, {"error": "not found"})
if WORKER_API_KEY and self.headers.get("x-sched-api-key") != WORKER_API_KEY:
return self._send(401, {"error": "unauthorized"})
run_id = self.headers.get("x-sched-run-id", "?")
try:
r = subprocess.run(WORKER_CMD, capture_output=True, text=True,
env=os.environ, timeout=WORKER_TIMEOUT)
except subprocess.TimeoutExpired:
return self._send(504, {"error": "worker timeout", "runId": run_id})
if r.returncode != 0:
return self._send(500, {"error": (r.stderr or r.stdout)[-500:], "runId": run_id})
try:
summary = json.loads(r.stdout)
except ValueError:
summary = {"stdout": r.stdout[-500:]}
self._send(200, {"ok": True, "runId": run_id, "summary": summary})
if __name__ == "__main__":
ThreadingHTTPServer(("0.0.0.0", PORT), Handler).serve_forever()