handling concurrency issues:

backend/app/plan/router.py

@@ -1,6 +1,9 @@
 import boto3
 import json
 import math
+import asyncio
+import random
+
 from datetime import datetime
 
 from fastapi import APIRouter, Depends
@@ -83,7 +86,6 @@ async def trigger_plan_entrypoint(body: PlanTriggerRequest):
                 index_end = min((i + 1) * chunk_size, total_rows)
 
                 message_payload = {**data, "index_start": index_start, "index_end": index_end}
-
                 message_body = json.dumps(message_payload)
 
                 response = sqs_client.send_message(
@@ -91,7 +93,10 @@ async def trigger_plan_entrypoint(body: PlanTriggerRequest):
                     MessageBody=message_body
                 )
                 logger.info(
-                    f"Chunk {i} sent to SQS. Rows {index_start}–{index_end}. Message ID: {response.get('MessageId')}")
+                    f"Chunk {i} sent to SQS. Rows {index_start}–{index_end}. Message ID: {response.get('MessageId')}"
+                )
+
+                await asyncio.sleep(random.uniform(0.1, 0.5))  # Delay to reduce spike pressure
 
         except Exception as e:
             logger.error("Error during Excel file handling: %s", e)

backend/ml_models/api.py

@@ -1,4 +1,5 @@
 import json
+import random
 import aiohttp
 import asyncio
 import pandas as pd
@@ -18,9 +19,6 @@ class ModelApi:
         "sap_change_predictions",
         "heat_demand_predictions",
         "carbon_change_predictions",
-        # "lighting_cost_predictions",
-        # "heating_cost_predictions",
-        # "hot_water_cost_predictions",
     ]
 
     KWH_MODEL_PREFIXES = ["heating_kwh_predictions", "hotwater_kwh_predictions"]
@@ -31,9 +29,6 @@ class ModelApi:
         "carbon_change_predictions": "carbonmodel",
         "hotwater_kwh_predictions": "hotwaterkwhmodel",
         "heating_kwh_predictions": "heatingkwhmodel",
-        # "lighting_cost_predictions": "lightingmodel",
-        # "heating_cost_predictions": "heatingmodel",
-        # "hot_water_cost_predictions": "hotwatermodel",
     }
 
     def __init__(
@@ -44,21 +39,12 @@ class ModelApi:
         base_url="https://api.dev.hestia.homes",
         max_retries=2,
     ):
-        """
-        This class handles the communication with the Model APIs. These models include SAP change, heat demain change
-        and carbon change
-
-        property_id (int, optional): :
-        :param portfolio_id: The portfolio ID to be passed in the request payload. Defaults to 4.
-        :param timestamp: The creation timestamp to be passed in the request payload. Defaults to None.
-        :param base_url:
-        """
         self.base_url = base_url
         self.portfolio_id = portfolio_id
         self.timestamp = timestamp
         self.prediction_buckets = prediction_buckets
         self.max_retries = max_retries
-        self.semaphore = asyncio.Semaphore(5)
+        self.semaphore = asyncio.Semaphore(2)
 
     @staticmethod
     def get_aiohttp_session():
@@ -131,6 +117,7 @@ class ModelApi:
         }
 
         async with self.semaphore:
+            await asyncio.sleep(random.uniform(0.3, 1.2))
             try:
                 async with session.post(url, json=payload, headers=headers, timeout=120) as response:
                     if response.status != 200:
@@ -215,18 +202,21 @@ class ModelApi:
     async def predict_all_async(self, df, bucket, model_prefixes=None, extract_ids=True) -> dict:
         model_prefixes = self.MODEL_PREFIXES if model_prefixes is None else model_prefixes
         predictions = {}
-        tasks = []
 
         session = self.get_aiohttp_session()
 
-        for model_prefix in model_prefixes:
+        async def run_model(model_prefix):
             logger.info(f"Scoring for model prefix: {model_prefix}")
             file_location = self.upload_scoring_data(df, bucket, model_prefix)
-            tasks.append(self.predict_async(f"s3://{bucket}/" + file_location, model_prefix, session=session))
+            response = await self.predict_async(f"s3://{bucket}/" + file_location, model_prefix, session=session)
+            return model_prefix, response
 
-        responses = await asyncio.gather(*tasks, return_exceptions=True)
+        results = []
+        for coro in asyncio.as_completed([run_model(mp) for mp in model_prefixes]):
+            result = await coro
+            results.append(result)
 
-        for model_prefix, response in zip(model_prefixes, responses):
+        for model_prefix, response in results:
             if response:
                 predictions_bucket = self.prediction_buckets[model_prefix]
                 predictions_df = pd.DataFrame(

serverless.yml

@@ -62,10 +62,13 @@ functions:
     timeout: 900
     memorySize: 3008
     role: EngineLambdaRole
+    reservedConcurrency: 5
     events:
       - sqs:
           arn: arn:aws:sqs:${self:provider.region}:${aws:accountId}:model-engine-queue
           batchSize: 1
+          maximumConcurrency: 2
+
 
 resources:
   Resources: