Merge pull request #289 from galgeek/metrics_plus_proxy_retries

prometheus metrics and yt-dlp external proxy retry updates

brozzler/__init__.py

@@ -39,6 +39,10 @@ class PageInterstitialShown(Exception):
     pass
 
 
+class VideoExtractorError(Exception):
+    pass
+
+
 class ProxyError(Exception):
     pass
 

brozzler/cli.py

@@ -234,6 +234,25 @@ def brozzle_page(argv=None):
         action="store_true",
         help="Try to avoid web bot detection",
     )
+    arg_parser.add_argument(
+        "--metrics_port",
+        type=int,
+        dest="metrics_port",
+        default=0,
+        help="Port for brozzler's Prometheus scrape endpoint",
+    )
+    arg_parser.add_argument(
+        "--registry_url",
+        dest="registry_url",
+        default=None,
+        help="http-sd-registry url, for Prometheus metrics discovery",
+    )
+    arg_parser.add_argument(
+        "--env",
+        dest="env",
+        default=None,
+        help="deployment environment for this brozzler instance, e.g., prod or qa",
+    )
     arg_parser.add_argument(
         "--screenshot-full-page", dest="screenshot_full_page", action="store_true"
     )
@@ -279,6 +298,9 @@ def brozzle_page(argv=None):
         window_height=args.window_height,
         window_width=args.window_width,
         stealth=args.stealth,
+        metrics_port=args.metrics_port,
+        registry_url=args.registry_url,
+        env=args.env,
     )
 
     def on_screenshot(screenshot_jpeg):
@@ -517,6 +539,25 @@ def brozzler_worker(argv=None):
         action="store_true",
         help="Try to avoid web bot detection",
     )
+    arg_parser.add_argument(
+        "--metrics_port",
+        type=int,
+        dest="metrics_port",
+        default=0,
+        help="Port for brozzler's Prometheus scrape endpoint",
+    )
+    arg_parser.add_argument(
+        "--registry_url",
+        dest="registry_url",
+        default=None,
+        help="http-sd-registry url, for Prometheus metrics discovery",
+    )
+    arg_parser.add_argument(
+        "--env",
+        dest="env",
+        default=None,
+        help="deployment environment for this brozzler instance, e.g., prod or qa",
+    )
     add_common_options(arg_parser, argv)
 
     args = arg_parser.parse_args(args=argv[1:])
@@ -573,6 +614,9 @@ def brozzler_worker(argv=None):
         skip_visit_hashtags=args.skip_visit_hashtags,
         skip_youtube_dl=args.skip_youtube_dl,
         stealth=args.stealth,
+        metrics_port=args.metrics_port,
+        registry_url=args.registry_url,
+        env=args.env,
     )
 
     signal.signal(signal.SIGQUIT, dump_state)

brozzler/metrics.py

@@ -0,0 +1,58 @@
+from typing import Optional
+
+try:
+    from http_sd_registry.client import (
+        Client,
+        Env,
+        Registration,
+        Scheme,
+        format_self_target,
+    )
+    from http_sd_registry.config import ClientConfig
+except ImportError:
+    # for users without access to http_sd_registry
+    http_sd_registry = None
+
+
+from prometheus_client import Counter, Gauge, Histogram, start_http_server
+
+# fmt: off
+brozzler_pages_crawled = Counter("brozzler_pages_crawled", "number of pages visited by brozzler")
+brozzler_page_processing_duration_seconds = Histogram("brozzler_page_processing_duration_seconds", "time spent processing a page in brozzler")
+brozzler_outlinks_found = Counter("brozzler_outlinks_found", "number of outlinks found by brozzler")
+brozzler_last_page_crawled_time = Gauge("brozzler_last_page_crawled_time", "time of last page visit, in seconds since UNIX epoch")
+brozzler_in_progress_pages = Gauge("brozzler_in_progress_pages", "number of pages currently processing with brozzler")
+brozzler_ydl_urls_checked = Counter("brozzler_ydl_urls_checked", "count of urls checked by brozzler yt-dlp")
+brozzler_ydl_extract_attempts = Counter("brozzler_ydl_extract_attempts", "count of extracts attempted by brozzler yt-dlp", labelnames=["youtube_host"])
+brozzler_ydl_extract_successes = Counter("brozzler_ydl_extract_successes", "count of extracts completed by brozzler yt-dlp", labelnames=["youtube_host"])
+brozzler_ydl_download_successes = Counter("brozzler_ydl_download_successes", "count of downloads completed by brozzler yt-dlp", labelnames=["youtube_host"])
+# fmt: on
+
+
+def register_prom_metrics(
+    metrics_port: int = 8888,
+    registry_url: Optional[str] = None,
+    env: Optional[str] = None,
+):
+    # Start metrics endpoint for scraping
+    start_http_server(metrics_port)
+
+    if registry_url is None:
+        return
+
+    if env == "qa":
+        env_for_prom = Env.qa
+    elif env == "prod":
+        env_for_prom = Env.prod
+    else:
+        env_for_prom = Env.qa
+
+    config = ClientConfig(server_url_base=registry_url)
+    client = Client(config)
+    target = format_self_target(scrape_port=metrics_port)
+    registration = Registration(
+        target=target,
+        env=env_for_prom,
+        scheme=Scheme.http,
+    )
+    client.keep_registered_threaded(registration)

brozzler/worker.py

@@ -35,6 +35,7 @@ import tempfile
 import urlcanon
 from requests.structures import CaseInsensitiveDict
 import rethinkdb as rdb
+from . import metrics
 from . import ydl
 
 r = rdb.RethinkDB()
@@ -71,6 +72,9 @@ class BrozzlerWorker:
         stealth=False,
         window_height=900,
         window_width=1400,
+        metrics_port=0,
+        registry_url=None,
+        env=None,
     ):
         self._frontier = frontier
         self._service_registry = service_registry
@@ -93,6 +97,9 @@ class BrozzlerWorker:
         self._window_height = window_height
         self._window_width = window_width
         self._stealth = stealth
+        self._metrics_port = metrics_port
+        self._registry_url = registry_url
+        self._env = env
 
         self._browser_pool = brozzler.browser.BrowserPool(
             max_browsers, chrome_exe=chrome_exe, ignore_cert_errors=True
@@ -104,6 +111,16 @@ class BrozzlerWorker:
         self._start_stop_lock = threading.Lock()
         self._shutdown = threading.Event()
 
+        # set up metrics
+        if self._metrics_port > 0:
+            metrics.register_prom_metrics(
+                self._metrics_port, self._registry_url, self._env
+            )
+        else:
+            logging.warning(
+                "not starting prometheus scrape endpoint: metrics_port is undefined"
+            )
+
     def _choose_warcprox(self):
         warcproxes = self._service_registry.available_services("warcprox")
         if not warcproxes:
@@ -266,6 +283,7 @@ class BrozzlerWorker:
             ):
                 try:
                     ydl_outlinks = ydl.do_youtube_dl(self, site, page)
+                    metrics.brozzler_ydl_urls_checked.inc(1)
                     outlinks.update(ydl_outlinks)
                 except brozzler.ReachedLimit as e:
                     raise
@@ -273,6 +291,11 @@ class BrozzlerWorker:
                     raise
                 except brozzler.ProxyError:
                     raise
+                except brozzler.VideoExtractorError as e:
+                    logging.error(
+                        "error extracting video info: %s",
+                        e,
+                    )
                 except Exception as e:
                     if (
                         hasattr(e, "exc_info")
@@ -311,7 +334,15 @@ class BrozzlerWorker:
             return False
         return True
 
+    @metrics.brozzler_page_processing_duration_seconds.time()
+    @metrics.brozzler_in_progress_pages.track_inprogress()
     def _browse_page(self, browser, site, page, on_screenshot=None, on_request=None):
+        def update_page_metrics(page, outlinks):
+            """Update page-level Prometheus metrics."""
+            metrics.brozzler_last_page_crawled_time.set_to_current_time()
+            metrics.brozzler_pages_crawled.inc(1)
+            metrics.brozzler_outlinks_found.inc(len(outlinks))
+
         def _on_screenshot(screenshot_jpeg):
             if on_screenshot:
                 on_screenshot(screenshot_jpeg)
@@ -416,6 +447,7 @@ class BrozzlerWorker:
         )
         if final_page_url != page.url:
             page.note_redirect(final_page_url)
+        update_page_metrics(page, outlinks)
         return outlinks
 
     def _fetch_url(self, site, url=None, page=None):

brozzler/ydl.py

@@ -27,10 +27,16 @@ import os
 import json
 import doublethink
 import datetime
+from . import metrics
 import threading
+import time
 
 thread_local = threading.local()
 
+YTDLP_PROXY = ""
+MAX_YTDLP_ATTEMPTS = 4
+YTDLP_WAIT = 10
+
 
 def should_ytdlp(site, page, page_status, skip_av_seeds):
     # called only after we've passed needs_browsing() check
@@ -65,6 +71,11 @@ def should_ytdlp(site, page, page_status, skip_av_seeds):
     return True
 
 
+def isyoutubehost(url):
+    # split 1 splits scheme from url, split 2 splits path from hostname, split 3 splits query string on hostname
+    return "youtube.com" in url.split("//")[-1].split("/")[0].split("?")[0]
+
+
 class ExtraHeaderAdder(urllib.request.BaseHandler):
     def __init__(self, extra_headers):
         self.extra_headers = extra_headers
@@ -191,15 +202,16 @@ def _build_youtube_dl(worker, destdir, site, page):
                     payload=f,
                     extra_headers=extra_headers,
                 )
-                # consulted by _remember_videos()
-                ydl.pushed_videos.append(
-                    {
-                        "url": url,
-                        "response_code": response.code,
-                        "content-type": mimetype,
-                        "content-length": size,
-                    }
-                )
+
+            # consulted by _remember_videos()
+            ydl.pushed_videos.append(
+                {
+                    "url": url,
+                    "response_code": response.code,
+                    "content-type": mimetype,
+                    "content-length": size,
+                }
+            )
 
     def maybe_heartbeat_site_last_claimed(*args, **kwargs):
         # in case yt-dlp takes a long time, heartbeat site.last_claimed
@@ -230,6 +242,9 @@ def _build_youtube_dl(worker, destdir, site, page):
             worker.logger.info(
                 "[ydl_postprocess_hook] postprocessor: {}".format(d["postprocessor"])
             )
+            is_youtube_host = isyoutubehost(d["info_dict"]["webpage_url"])
+
+            metrics.brozzler_ydl_download_successes.labels(is_youtube_host).inc(1)
             if worker._using_warcprox(site):
                 _YoutubeDL._push_video_to_warcprox(
                     _YoutubeDL, site, d["info_dict"], d["postprocessor"]
@@ -262,9 +277,22 @@ def _build_youtube_dl(worker, destdir, site, page):
         "logger": logging.getLogger("yt_dlp"),
         "verbose": False,
         "quiet": False,
+        # recommended to avoid bot detection
+        "sleep_interval": 25,
+        "max_sleep_interval": 90,
     }
 
-    # skip proxying yt-dlp v.2023.07.06
+    ytdlp_url = page.redirect_url if page.redirect_url else page.url
+    is_youtube_host = isyoutubehost(ytdlp_url)
+    if is_youtube_host and YTDLP_PROXY:
+        ydl_opts["proxy"] = YTDLP_PROXY
+        # don't log proxy value secrets
+        ytdlp_proxy_for_logs = (
+            YTDLP_PROXY.split("@")[1] if "@" in YTDLP_PROXY else "@@@"
+        )
+        logging.info("using yt-dlp proxy ... %s", ytdlp_proxy_for_logs)
+
+    # skip warcprox proxying yt-dlp v.2023.07.06: youtube extractor using ranges
     # if worker._proxy_for(site):
     #    ydl_opts["proxy"] = "http://{}".format(worker._proxy_for(site))
 
@@ -272,6 +300,8 @@ def _build_youtube_dl(worker, destdir, site, page):
     if site.extra_headers():
         ydl._opener.add_handler(ExtraHeaderAdder(site.extra_headers(page)))
     ydl.pushed_videos = []
+    ydl.url = ytdlp_url
+    ydl.is_youtube_host = is_youtube_host
 
     return ydl
 
@@ -295,58 +325,85 @@ def _remember_videos(page, pushed_videos=None):
 
 
 def _try_youtube_dl(worker, ydl, site, page):
-    ytdlp_url = page.redirect_url if page.redirect_url else page.url
-    try:
-        logging.info("trying yt-dlp on %s", ytdlp_url)
-
-        with brozzler.thread_accept_exceptions():
-            # we do whatwg canonicalization here to avoid "<urlopen error
-            # no host given>" resulting in ProxyError
-            # needs automated test
-            # and yt-dlp needs sanitize_info for extract_info
-            ie_result = ydl.sanitize_info(
-                ydl.extract_info(str(urlcanon.whatwg(ytdlp_url)))
-            )
-        _remember_videos(page, ydl.pushed_videos)
-        if worker._using_warcprox(site):
-            info_json = json.dumps(ie_result, sort_keys=True, indent=4)
-            logging.info(
-                "sending WARCPROX_WRITE_RECORD request to warcprox "
-                "with yt-dlp json for %s",
-                ytdlp_url,
-            )
-            worker._warcprox_write_record(
-                warcprox_address=worker._proxy_for(site),
-                url="youtube-dl:%s" % str(urlcanon.semantic(ytdlp_url)),
-                warc_type="metadata",
-                content_type="application/vnd.youtube-dl_formats+json;charset=utf-8",
-                payload=info_json.encode("utf-8"),
-                extra_headers=site.extra_headers(page),
-            )
-        return ie_result
-    except brozzler.ShutdownRequested as e:
-        raise
-    except Exception as e:
-        if hasattr(e, "exc_info") and e.exc_info[0] == yt_dlp.utils.UnsupportedError:
-            return None
-        elif (
-            hasattr(e, "exc_info")
-            and e.exc_info[0] == urllib.error.HTTPError
-            and hasattr(e.exc_info[1], "code")
-            and e.exc_info[1].code == 420
-        ):
-            raise brozzler.ReachedLimit(e.exc_info[1])
-        elif (
-            hasattr(e, "exc_info")
-            and e.exc_info[0] == urllib.error.URLError
-            and worker._proxy_for(site)
-        ):
-            # connection problem when using a proxy == proxy error (XXX?)
-            raise brozzler.ProxyError(
-                "yt-dlp hit apparent proxy error from " "%s" % ytdlp_url
-            ) from e
-        else:
+    attempt = 0
+    while attempt < MAX_YTDLP_ATTEMPTS:
+        try:
+            logging.info("trying yt-dlp on %s", ydl.url)
+            # should_download_vid = not ydl.is_youtube_host
+            # then
+            # ydl.extract_info(str(urlcanon.whatwg(ydl.url)), download=should_download_vid)
+            # if ydl.is_youtube_host and ie_result:
+            #     download_url = ie_result.get("url")
+            metrics.brozzler_ydl_extract_attempts.labels(ydl.is_youtube_host).inc(1)
+            with brozzler.thread_accept_exceptions():
+                # we do whatwg canonicalization here to avoid "<urlopen error
+                # no host given>" resulting in ProxyError
+                # needs automated test
+                # and yt-dlp needs sanitize_info for extract_info
+                ie_result = ydl.sanitize_info(
+                    ydl.extract_info(str(urlcanon.whatwg(ydl.url)))
+                )
+            metrics.brozzler_ydl_extract_successes.labels(ydl.is_youtube_host).inc(1)
+            break
+        except brozzler.ShutdownRequested as e:
             raise
+        except Exception as e:
+            if (
+                hasattr(e, "exc_info")
+                and e.exc_info[0] == yt_dlp.utils.UnsupportedError
+            ):
+                return None
+            elif (
+                hasattr(e, "exc_info")
+                and e.exc_info[0] == urllib.error.HTTPError
+                and hasattr(e.exc_info[1], "code")
+                and e.exc_info[1].code == 420
+            ):
+                raise brozzler.ReachedLimit(e.exc_info[1])
+            else:
+                # todo: other errors to handle separately?
+                # OSError('Tunnel connection failed: 464 Host Not Allowed') (caused by ProxyError...)
+                # and others...
+                attempt += 1
+                if attempt == MAX_YTDLP_ATTEMPTS:
+                    logging.warning(
+                        "Failed after %s attempts. Error: %s", MAX_YTDLP_ATTEMPTS, e
+                    )
+                    raise brozzler.VideoExtractorError(
+                        "yt-dlp hit error extracting info for %s" % ydl.url
+                    )
+                else:
+                    retry_wait = min(60, YTDLP_WAIT * (1.5 ** (attempt - 1)))
+                    logging.info(
+                        "Attempt %s failed. Retrying in %s seconds...",
+                        attempt,
+                        retry_wait,
+                    )
+                    time.sleep(retry_wait)
+    else:
+        raise brozzler.VideoExtractorError(
+            "yt-dlp hit unknown error extracting info for %s" % ydl.url
+        )
+
+    logging.info("ytdlp completed successfully")
+
+    _remember_videos(page, ydl.pushed_videos)
+    if worker._using_warcprox(site):
+        info_json = json.dumps(ie_result, sort_keys=True, indent=4)
+        logging.info(
+            "sending WARCPROX_WRITE_RECORD request to warcprox "
+            "with yt-dlp json for %s",
+            ydl.url,
+        )
+        worker._warcprox_write_record(
+            warcprox_address=worker._proxy_for(site),
+            url="youtube-dl:%s" % str(urlcanon.semantic(ydl.url)),
+            warc_type="metadata",
+            content_type="application/vnd.youtube-dl_formats+json;charset=utf-8",
+            payload=info_json.encode("utf-8"),
+            extra_headers=site.extra_headers(page),
+        )
+    return ie_result
 
 
 def do_youtube_dl(worker, site, page):

setup.py

@@ -34,7 +34,7 @@ def find_package_data(package):
 
 setuptools.setup(
     name="brozzler",
-    version="1.5.54",
+    version="1.5.55",
     description="Distributed web crawling with browsers",
     url="https://github.com/internetarchive/brozzler",
     author="Noah Levitt",
@@ -77,9 +77,10 @@ setuptools.setup(
         "jinja2>=2.10",
         "cryptography>=2.3",
         "python-magic>=0.4.15",
+        "prometheus-client>=0.20.0",
     ],
     extras_require={
-        "yt-dlp": ["yt-dlp==2024.7.25"],
+        "yt-dlp": ["yt-dlp>=2024.7.25"],
         "dashboard": ["flask>=1.0", "gunicorn>=19.8.1"],
         "easy": [
             "warcprox>=2.4.31",