Track autoPausedEmailsSent state in db instead of crawl state

This is much more reliable, prevents duplicate emails as was
sometimes happening before, and makes it easier to clear the state
when a crawl is unpaused.

Author: tw4l

backend/btrixcloud/crawls.py

@@ -838,6 +838,11 @@ async def pause_crawl(
         if pause and not paused_at:
             paused_at = dt_now()
 
+        if not pause:
+            # If unpausing, unset autoPausedEmailsSent so that we will send
+            # emails again if quota is reached
+            await self.set_auto_paused_emails_sent(crawl_id, org, False)
+
         try:
             result = await self.crawl_manager.pause_resume_crawl(
                 crawl_id, paused_at=paused_at
@@ -1210,6 +1215,7 @@ async def get_crawl_logs(
     async def notify_org_admins_of_auto_paused_crawl(
         self,
         paused_reason: TYPE_AUTO_PAUSED_STATES,
+        crawl_id: str,
         cid: UUID,
         org: Organization,
     ):
@@ -1231,6 +1237,29 @@ async def notify_org_admins_of_auto_paused_crawl(
             ]
         )
 
+        await self.set_auto_paused_emails_sent(crawl_id, org)
+
+    async def set_auto_paused_emails_sent(
+        self, crawl_id: str, org: Organization, emails_sent: bool = True
+    ):
+        """Set if auto-paused emails already sent"""
+        await self.crawls.find_one_and_update(
+            {"_id": crawl_id, "oid": org.id, "type": "crawl"},
+            {"$set": {"autoPausedEmailsSent": emails_sent}},
+        )
+
+    async def get_auto_paused_emails_sent(
+        self, crawl_id: str, org: Organization
+    ) -> bool:
+        """Return whether auto-paused emails already sent for crawl"""
+        res = await self.crawls.find_one(
+            {"_id": crawl_id, "oid": org.id, "type": "crawl"},
+            projection=["autoPausedEmailsSent"],
+        )
+        if res:
+            return res.get("autoPausedEmailsSent", False)
+        return False
+
 
 # ============================================================================
 async def recompute_crawl_file_count_and_size(crawls, crawl_id: str):

backend/btrixcloud/models.py

@@ -1092,6 +1092,8 @@ class Crawl(BaseCrawl, CrawlConfigCore):
 
     pendingSize: int = 0
 
+    autoPausedEmailsSent: bool = False
+
 
 # ============================================================================
 class CrawlCompleteIn(BaseModel):

backend/btrixcloud/operator/crawls.py

@@ -1654,13 +1654,18 @@ async def update_crawl_state(
                     allowed_from=RUNNING_AND_WAITING_STATES,
                 )
 
-                if paused_state != "paused" and not status.autoPausedEmailsSent:
+                if (
+                    paused_state != "paused"
+                    and not await self.crawl_ops.get_auto_paused_emails_sent(
+                        crawl.id, crawl.org
+                    )
+                ):
                     await self.crawl_ops.notify_org_admins_of_auto_paused_crawl(
                         paused_reason=paused_state,
+                        crawl_id=crawl.id,
                         cid=crawl.cid,
                         org=crawl.org,
                     )
-                    status.autoPausedEmailsSent = True
 
                 return status
 

backend/btrixcloud/operator/models.py

@@ -272,6 +272,3 @@ class CrawlStatus(BaseModel):
 
     # last state
     last_state: TYPE_ALL_CRAWL_STATES = Field(default="starting", exclude=True)
-
-    # email sent to org admins because crawl was auto-paused
-    autoPausedEmailsSent: bool = False