temporarily disable calling check_tw_to_close() in redis per flow to see if its the main source of latency

Author: AlyaGomaa

slips_files/core/database/redis_db/profile_handler.py

@@ -1577,8 +1577,9 @@ def mark_profile_tw_as_modified(self, profileid, twid, timestamp):
         data = {f"{profileid}{self.separator}{twid}": float(timestamp)}
         self.r.zadd(self.constants.MODIFIED_TIMEWINDOWS, data)
         self.publish("tw_modified", f"{profileid}:{twid}")
-        # Check if we should close some TW
-        self.check_tw_to_close()
+        # # Check if we should close some TW
+        # self.check_tw_to_close() # @@@@@@@@@@@@@@@@@ TODO temporarily
+        #  disabled
 
     def publish_new_letter(
         self, new_symbol: str, profileid: str, twid: str, tupleid: str, flow

slips_files/core/profiler_worker.py

@@ -285,6 +285,7 @@ def store_features_going_out(self, flow, profileid, twid):
 
         # now that slips successfully parsed the flow,
         # mark this profile as modified
+        # @@@@@@@@@@@@@ !! TODO this is the second suspect,
         self.db.mark_profile_tw_as_modified(profileid, twid, "")
         return True