feat: track docs that fail pdfize task

ietf/doc/migrations/0022_unprocessabledocument.py

@@ -0,0 +1,46 @@
+# Generated by Django 4.2.14 on 2024-07-25 22:28
+
+from django.db import migrations, models
+import django.db.models.deletion
+import django.utils.timezone
+import ietf.utils.models
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ("doc", "0021_narrativeminutes"),
+    ]
+
+    operations = [
+        migrations.CreateModel(
+            name="UnprocessableDocument",
+            fields=[
+                (
+                    "id",
+                    models.AutoField(
+                        auto_created=True,
+                        primary_key=True,
+                        serialize=False,
+                        verbose_name="ID",
+                    ),
+                ),
+                ("rev", models.CharField(max_length=16, verbose_name="revision")),
+                (
+                    "proc_type",
+                    models.CharField(
+                        choices=[("pdfize", "Pdfize")],
+                        help_text="type of processing that fails for this document",
+                        max_length=16,
+                    ),
+                ),
+                ("time", models.DateTimeField(default=django.utils.timezone.now)),
+                (
+                    "document",
+                    ietf.utils.models.ForeignKey(
+                        on_delete=django.db.models.deletion.CASCADE, to="doc.document"
+                    ),
+                ),
+            ],
+        ),
+    ]

ietf/doc/models.py

@@ -1483,3 +1483,23 @@ class BofreqEditorDocEvent(DocEvent):
 class BofreqResponsibleDocEvent(DocEvent):
     """ Capture the responsible leadership (IAB and IESG members) for a BOF Request """
     responsible = models.ManyToManyField('person.Person', blank=True)
+
+
+
+class UnprocessableDocument(models.Model):
+    """Document """
+    class ProcTypes(models.TextChoices):
+        PDFIZE = "pdfize"
+
+    document = ForeignKey(Document)
+    rev = models.CharField(
+        verbose_name="revision",
+        max_length=16,
+        blank=False,
+    )
+    proc_type = models.CharField(
+        max_length=16,
+        choices=ProcTypes.choices,
+        help_text="type of processing that fails for this document",
+    )
+    time = models.DateTimeField(default=timezone.now)

ietf/doc/tasks.py

@@ -12,6 +12,7 @@
 from django.conf import settings
 from django.utils import timezone
 
+from ietf.doc.models import UnprocessableDocument
 from ietf.utils import log
 from ietf.utils.celery import celery_task_lock
 from ietf.utils.timezone import datetime_today
@@ -122,15 +123,15 @@ def generate_draft_bibxml_files_task(days=7, process_all=False):
             log.log(f"Error generating bibxml for {event.doc.name}-{event.rev}: {err}")
 
 
-@shared_task(bind=True, time_limit=30, soft_time_limit=28)
+@shared_task(bind=True, time_limit=32, soft_time_limit=30)
 def pdfize_document_task(self, name, rev):
     doc = Document.objects.filter(name=name).first()
     if doc is None:
         log.log(f"Failed to pdfize document {name} rev {rev}: Document does not exist")
         return
     # There is a very slight race condition between the task time_limit and the lock expiration
     # time. We can tolerate this task running twice if that unlikely timing ever works out.
-    with celery_task_lock(self, 30) as acquired:
+    with celery_task_lock(self, 32) as acquired:
         if not acquired:
             log.log(f"{self.name}({name}, {rev}) skipped because it's already running")
             return
@@ -140,3 +141,11 @@ def pdfize_document_task(self, name, rev):
             PdfizedDoc(doc).update_cache()
         except SoftTimeLimitExceeded:
             log.log(f"Failed to pdfize document {name} rev {rev}: exceeded task time limit")
+            UnprocessableDocument.objects.update_or_create(
+                document=doc,
+                rev=doc.rev,
+                proc_type=UnprocessableDocument.ProcTypes.PDFIZE,
+                defaults={
+                    "time": timezone.now(),
+                },
+            )

ietf/doc/views_doc.py

@@ -57,7 +57,7 @@
 from ietf.doc.models import ( Document, DocHistory, DocEvent, BallotDocEvent, BallotType,
     ConsensusDocEvent, NewRevisionDocEvent, TelechatDocEvent, WriteupDocEvent, IanaExpertDocEvent,
     IESG_BALLOT_ACTIVE_STATES, STATUSCHANGE_RELATIONS, DocumentActionHolder, DocumentAuthor,
-    RelatedDocument, RelatedDocHistory)
+    RelatedDocument, RelatedDocHistory, UnprocessableDocument)
 from ietf.doc.utils import (augment_events_with_revision,
     can_adopt_draft, can_unadopt_draft, get_chartering_type, get_tags_for_stream_id, investigate_fragment,
     needed_ballot_positions, nice_consensus, update_telechat, has_same_ballot,
@@ -1067,7 +1067,11 @@ def document_pdfized(request, name, rev=None, ext=None):
     pdf = PdfizedDoc(doc).get()
     if pdf:
         return HttpResponse(pdf, content_type="application/pdf")
-
+    elif UnprocessableDocument.objects.filter(
+        document=doc, rev=doc.rev, proc_type=UnprocessableDocument.ProcTypes.PDFIZE
+    ).exists():
+        return Http404()
+
     pdfize_document_task.delay(name=doc.name, rev=doc.rev)
     refresh_time_seconds = 5
     path=request.path