T-HUB/infra/peertube-workers/observe.cjs

'use strict';
// Transfer observation and canonical-origin routing. PeerTube retains queue selection and retries.
const fs = require('node:fs');
const path = require('node:path');
const { AsyncLocalStorage } = require('node:async_hooks');
const context = new AsyncLocalStorage();
const state = process.env.PT_STATE_DIR || '/var/lib/pt-worker/state';
let route;
try {
  const cfg = JSON.parse(fs.readFileSync(process.env.PT_WORKER_CONFIG || '/etc/pt-worker/worker.json'));
  if (cfg.canonical_origin && cfg.peertube_url) route = { from: new URL(cfg.canonical_origin), to: new URL(cfg.peertube_url) };
} catch { /* Enrollment and isolated tests may not yet have a config. */ }
const originals = { 'http:': require('node:http').request, 'https:': require('node:https').request };
function emit(data) {
  try {
    fs.mkdirSync(state, { recursive: true, mode: 0o700 });
    fs.appendFileSync(path.join(state, 'events.jsonl'), JSON.stringify({ time: Date.now() / 1000, ...data }) + '\n', { mode: 0o600 });
  } catch { /* Observation never changes job behavior. */ }
}
const createWriteStream = fs.createWriteStream;
fs.createWriteStream = function (filename, ...args) {
  const pathname = context.getStore();
  const match = pathname?.match(/\/api\/v1\/runners\/jobs\/([a-f\d-]{36})\/files\//);
  if (match && typeof filename === 'string') {
    try {
      const dir = path.join(state, 'inputs');
      fs.mkdirSync(dir, { recursive: true, mode: 0o700 });
      fs.writeFileSync(path.join(dir, path.basename(filename) + '.json'), JSON.stringify({ uuid: match[1] }), { mode: 0o600 });
    } catch { /* Optional correlation only. */ }
  }
  return createWriteStream.call(this, filename, ...args);
};
require('node:module').syncBuiltinESMExports();
for (const name of ['node:http', 'node:https']) {
  const mod = require(name);
  const original = mod.request;
  mod.request = function (...args) {
    let selected = original;
    // Source payloads use the public origin. Route only that origin via Tailscale.
    let options = args[0];
    if (typeof options === 'string' || options instanceof URL) {
      const u = new URL(options);
      options = { protocol: u.protocol, hostname: u.hostname, port: u.port, path: u.pathname + u.search,
        ...(typeof args[1] === 'object' ? args[1] : {}) };
    }
    if (route && (options.hostname || options.host) === route.from.hostname &&
        (options.protocol || (name === 'node:https' ? 'https:' : 'http:')) === route.from.protocol &&
        String(options.port || (route.from.protocol === 'https:' ? 443 : 80)) === String(route.from.port || (route.from.protocol === 'https:' ? 443 : 80))) {
      const callback = typeof args[args.length - 1] === 'function' ? args[args.length - 1] : undefined;
      options = { ...options, protocol: route.to.protocol, hostname: route.to.hostname, port: route.to.port,
        headers: { ...(options.headers || {}), host: route.from.host } };
      delete options.host;
      selected = originals[route.to.protocol];
      args = callback ? [options, callback] : [options];
    }
    let req;
    if (typeof args[args.length - 1] === 'function') {
      const callback = args[args.length - 1];
      args[args.length - 1] = function (...values) {
        return context.run((req.path || '').split('?')[0], () => callback.apply(this, values));
      };
    }
    req = selected.apply(this, args);
    // No host, URL, query, headers, body, or tokens are ever recorded.
    const pathname = (req.path || '').split('?')[0];
    const match = pathname.match(/\/api\/v1\/runners\/jobs\/([a-f\d-]{36})\/(accept|success|error|abort)/);
    const download = pathname.includes('/files/');
    const upload = (match && match[2] === 'success') || (req.method === 'PUT');
    const begun = performance.now();
    let sent = 0, received = 0, done = false, requestErrored = false;
    function size(value, encoding) {
      if (Buffer.isBuffer(value) || value instanceof Uint8Array) return value.length;
      if (typeof value === 'string') return Buffer.byteLength(value, typeof encoding === 'string' ? encoding : 'utf8');
      return 0;
    }
    const write = req.write;
    req.write = function (chunk, ...rest) { sent += size(chunk, rest[0]); return write.call(this, chunk, ...rest); };
    const end = req.end;
    req.end = function (chunk, ...rest) { sent += size(chunk, rest[0]); return end.call(this, chunk, ...rest); };
    function finish(ok, code) {
      if (done) return;
      done = true;
      const seconds = Math.max(0.001, (performance.now() - begun) / 1000);
      if (download || upload) emit({ kind: 'transfer', direction: download ? 'download' : 'upload',
        bytes: download ? received : sent, seconds, ok, status_code: code || null });
      if (match) emit({ kind: 'job_' + match[2], uuid: match[1], ok });
    }
    req.on('response', res => {
      function interrupted() {
        finish(false, res.statusCode);
        req.destroy();
        // A ClientRequest may already be marked destroyed when its response aborts;
        // destroy(error) then emits nothing. Forward one error to the upstream promise.
        queueMicrotask(() => {
          if (!requestErrored) req.emit('error', new Error('Worker transfer stream interrupted'));
        });
      }
      res.on('data', chunk => { received += chunk.length; });
      res.once('end', () => finish(res.statusCode >= 200 && res.statusCode < 300, res.statusCode));
      res.once('aborted', interrupted);
      res.once('error', interrupted);
    });
    req.on('error', () => { requestErrored = true; finish(false); });
    return req;
  };
}
require('node:module').syncBuiltinESMExports();