Make the readiness check catch fetching that is scheduled but failing (#27)

/health/ready/ proved beat and a worker were alive but not that
fetching worked: every feed could be failing, or the queue could be
hours behind, and it would still report ok.

apps.core.health adds three checks: Celery queue depth (real Redis list
length, over 1000 fails), a successful fetch within three poll intervals
(30 min minimum, with a grace period for a brand-new install), and the
share of active feeds failing (over half, once there are at least four).
The payload also reports queue depth, active/failing feed counts, and
the age of the last successful fetch. Verified against real Redis with
queued Celery tasks, and the suite on PostgreSQL. docs/install.md lists
each check.

Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
This commit is contained in:
ksmithandClaude Sonnet 5.5 committed 2026-11-01 09:30:00 -07:00
1 parent f8cd1eb33e
commit fd494a9c08
4 files changed
+252 -11

No files matched your search

+83
View File
@@ -0,0 +1,83 @@
from datetime import timedelta
import redis
from django.conf import settings
from django.db.models import Max, Min
from django.utils import timezone
from apps.feeds.models import Feed
from .models import SiteSettings
# More than this many tasks waiting means workers aren't keeping up (or
# aren't running) — fetches are piling up faster than they complete.
MAX_QUEUE_DEPTH = 1000
# Past this share of active feeds failing, it's far more likely something
# on our side (DNS, outbound network, a bad deploy) than every publisher
# breaking at once. Needs a handful of feeds first so one broken feed out
# of two doesn't read as an outage.
MAX_FAILING_RATIO = 0.5
MIN_FEEDS_FOR_RATIO = 4
MIN_SUCCESS_WINDOW_MINUTES = 30
def queue_depth() -> int | None:
"""Tasks waiting in Celery's default queue, or None if the broker
can't be reached. Celery's Redis transport keeps each queue as a Redis
list named after the queue."""
try:
client = redis.Redis.from_url(
settings.CELERY_BROKER_URL, socket_connect_timeout=2, socket_timeout=2
)
return int(client.llen("celery"))
except Exception:
return None
def _success_window() -> timedelta:
floor = getattr(settings, "EVECTION_MIN_FETCH_INTERVAL_MINUTES", 15)
base = max(SiteSettings.load().fetch_interval_minutes, floor)
return timedelta(minutes=max(3 * base, MIN_SUCCESS_WINDOW_MINUTES))
def fetch_pipeline_status(now=None) -> tuple[dict[str, bool], dict]:
"""Checks that fetching is actually *succeeding*, not just scheduled
(the poll heartbeat only proves beat and a worker are alive): returns
({check: passed}, detail) covering queue depth, whether any feed has
been fetched successfully recently, and the share of feeds failing.
"""
now = now or timezone.now()
detail: dict = {}
depth = queue_depth()
detail["queue_depth"] = depth
checks = {"queue_depth": depth is not None and depth <= MAX_QUEUE_DEPTH}
active = Feed.objects.filter(is_active=True)
active_count = active.count()
failing = active.filter(fetch_error_count__gt=0).count()
detail["active_feeds"] = active_count
detail["failing_feeds"] = failing
window = _success_window()
last_success = active.filter(
last_fetch_status__in=[Feed.STATUS_OK, Feed.STATUS_NOT_MODIFIED]
).aggregate(latest=Max("last_fetched_at"))["latest"]
if last_success is not None:
detail["last_successful_fetch_age_seconds"] = int((now - last_success).total_seconds())
recent_success = now - last_success <= window
elif active_count == 0:
recent_success = True
else:
# Nothing has ever fetched successfully — only a problem once
# the oldest feed has had a full window to be fetched.
oldest = active.aggregate(oldest=Min("created_at"))["oldest"]
recent_success = now - oldest <= window
checks["recent_fetch_success"] = recent_success
checks["feed_failures"] = not (
active_count >= MIN_FEEDS_FOR_RATIO and failing / active_count > MAX_FAILING_RATIO
)
return checks, detail