diff --git a/merge_tool.py b/merge_tool.py
new file mode 100644
index 0000000..64294bf
--- /dev/null
+++ b/merge_tool.py
@@ -0,0 +1,9 @@
+from logzero import logger
+
+from utils import merge_lag_faults
+
+def main():
+    merge_lag_faults()
+
+if __name__ == "__main__":
+    main()
diff --git a/utils.py b/utils.py
index 881538e..8b1b864 100644
--- a/utils.py
+++ b/utils.py
@@ -1,10 +1,12 @@
 import json
 import os
 import re
+import pandas as pd
 from datetime import datetime as dtime
 from datetime import timedelta
 from typing import Dict, List
 from urllib.parse import urlparse
+from sqlalchemy import create_engine
 
 import marko
 import pytz
@@ -22,6 +24,7 @@
 CLEANR = re.compile("<.*?>")
 
 FAULT_RECORD_API_URL = os.getenv("FAULT_RECORD_API_URL")
+FAULT_RECORD_DB_URI = os.getenv("FAULT_RECORD_DB_URI")
 
 SHIB_MEMBER = os.getenv("SHIB_MEMBER")
 SHIB_FIRST_NAME = os.getenv("SHIB_FIRST_NAME")
@@ -30,6 +33,78 @@
 
 SCRAPER_USER_ID = -1
 
+def merge_lag_faults():
+
+    global SCRAPER_USER_ID
+
+    if SCRAPER_USER_ID == -1:
+        SCRAPER_USER_ID = get_or_create_user(SCRAPER_EMAIL, SHIB_FIRST_NAME, SHIB_LAST_NAME, FAULT_RECORD_API_URL)
+
+    FRengine = create_engine(FAULT_RECORD_DB_URI)
+    query = f"""
+        SELECT
+            *
+        FROM fault_record.records
+        """
+    df = pd.read_sql(query , FRengine)
+
+    # Scrape information needed to merge
+    regex = r"\*(.*)\* is (\d+) days old - \(last update: (.*)\)  (.*)"
+    df['split_desc'] = df.desc.apply(lambda x: re.findall(regex, x))
+    cols = ['scraped_source', 'lag', 'date', 'scraped_desc']
+    regex = r"\*(?P<scraped_source>.*)\* is (?P<lag>\d+) days old - \(last update: (?P<date>.*)\)  (?P<scraped_desc>.*)"
+    regex_cols = df.desc.str.extract(regex)
+    for col in cols:
+        df[col] = regex_cols[col]
+
+    # Drop the cols that don't have the matching data
+    df = df.drop('DVC_id', axis=1).dropna()
+    df['lag'] = df.lag.astype(int)
+
+    # Aggregate the faults we want to merge
+    df_merged = df.groupby(['scraped_source', 'date'])\
+        .agg({
+            'first_occurance': 'min',
+            'last_occurance': 'max',
+            'record_date': 'min',
+            'fault_id': 'min',
+            'lag': ['max', pd.Series.nunique]
+    }).reset_index()
+
+    df_merged.columns = ['scraped_source', 'date', 'first_occurance', 'last_occurance',
+                        'record_date', 'fault_id', 'lag', 'rows']
+
+    
+    headers = {"Content-type": "application/json", "Accept": "text/plain", "Member": SHIB_MEMBER, "givenName": SHIB_FIRST_NAME, "sn": SHIB_LAST_NAME}
+
+    # Update the first fault with the range of data
+    for _, row in df_merged.iterrows():
+        if row['rows'] > 1:
+            desc = df[df['fault_id']==row['fault_id']]['desc'].iloc[0]
+            url = f"{FAULT_RECORD_API_URL}/api/v1/admin/faults/{row['fault_id']}"
+            payload = {'published': '', 'user_id': SCRAPER_USER_ID}
+            payload['name'] = f"Merged lag faults for {row['scraped_source']}"
+            payload['desc'] = f"Merged faults with lag that grew to {row['lag']} with messages similar to \n {desc}"
+            payload['first_occurance'] = row['first_occurance'].isoformat()
+            payload['last_occurance'] = row['last_occurance'].isoformat()
+            payload['record_date'] = row['record_date'].isoformat()
+        
+            logger.info(f"Merging {row['rows']} records into fault# {row['fault_id']}.")
+            r = requests.put(url, data=json.dumps(payload), headers=headers)
+
+            if not r.ok:
+                logger.error(f"Something went wrong when trying to merge fault# {row['fault_id']}: {r.json()}")
+
+    # Delete the faults that were merged into the first
+    df_delete = df['fault_id'][~df['fault_id'].isin(df_merged['fault_id'])]
+    for fault_id in df_delete:
+        url = f"{FAULT_RECORD_API_URL}/api/v1/admin/faults/{fault_id}"
+
+        logger.info(f"Deleting fault# {fault_id}.")
+        r = requests.delete(url, headers=headers)
+        
+        if not r.ok:
+            logger.error(f"Something went wrong when trying to delete fault# {fault_id}: {r.json()}")
 
 def parse_timestamp(ts: float) -> str:
     """Convert Slack message timestamp to date
@@ -254,7 +329,7 @@ def get_signal_ids(message: str) -> List:
     """
     try:
         source, signals = extract_source_signal_pair(message)
-        query_signals_url = get_signals_url(source, signals)
+        query_signals_url = get_signals_url(source, FAULT_RECORD_API_URL, signals)
         signals = requests.get(query_signals_url)
         signal_ids = [sig.get("signal_id") for sig in signals.json()]
         return signal_ids