Added resources, created a prototype trigger and its corresponding dockerfile for update configurations, Testing required

clickhouse/config_update_scripts/Dockerfile

@@ -0,0 +1,9 @@
+FROM python
+
+WORKDIR /update_scripts
+
+COPY . /update_scripts/
+
+RUN pip install --no-cache-dir -r requirements.txt
+
+CMD ["python3","update_trigger.py"]

clickhouse/config_update_scripts/requirements.txt

@@ -0,0 +1,8 @@
+subprocess
+json
+jinja2
+yaml
+re
+xml
+schedule
+time

clickhouse/config_update_scripts/update_compose.py

@@ -4,6 +4,7 @@ import subprocess
 import json
 import xml.etree.ElementTree as ET
 import os
+import re
 
 if __name__ == "__main__":
     
@@ -15,14 +16,10 @@ if __name__ == "__main__":
 
     all_services = [json.loads(s) for s in all_services]
     # extracting the name, removing the custom id from it and storing it in a list
-    all_service_names = [service['Names'].split('.')[0] for service in all_services]
-    # extracting only 'keeper1', 'server1'...
+    all_service_names = [service['Names'].split('.')[0] for service in all_services if re.findall(r'clickhouse-server',service['Names'])]
+    # extracting only 'server1','server2'...
     all_service_names = [ name.split('-')[-1] for name in all_service_names]
 
-    # removing all keeepers
-    all_service_names.remove('keeper1')
-    all_service_names.remove('keeper2')
-    all_service_names.remove('keeper3')
     curr_num_servers = sorted(all_service_names)[-1][-1]
 
     replication_factor = 2

clickhouse/config_update_scripts/update_trigger.py

@@ -0,0 +1,34 @@
+import subprocess
+import json
+import re
+import schedule
+import time
+
+def check_util_exec():
+    # extracting details of each running container in json format    
+    try:
+        all_services = subprocess.check_output(["docker","stats","--no-stream","--format","json"],text=True).split('\n')[:-1]
+    except subprocess.CalledProcessError as e:
+        print(f"Command failed with return code {e.returncode}")
+
+    all_services = [json.loads(s) for s in all_services]
+
+    resource_util_exceed_flag = True            # Flag to check if all of the containers have exceeded 80% memory utilization 
+    for service in all_services:
+        if re.findall(r'clickhouse-server',service['Name']):
+            if float(service['MemPerc'][:-1]) < 80:
+                resource_util_exceed_flag = False
+    
+    if resource_util_exceed_flag:
+        process = subprocess.Popen(['python3','update_compose.py'],text=True,stdout=subprocess.PIPE,stderr=subprocess.PIPE)
+        stdout, stderr = process.communicate()  # Wait for the process to finish and capture output
+        print("Standard Output:", stdout)
+        print("Standard Error:", stderr)
+
+if __name__ == "__main__":
+    schedule.every(30).seconds.do(check_util_exec)
+    while True:
+        schedule.run_pending()
+        time.sleep(1)
+
+        

clickhouse/docker-compose.yaml

@@ -1,4 +1,9 @@
 services:
+  registry:
+    image: registry:2
+    ports:
+      - "5000:5000"
+      
   clickhouse-keeper1:
     image: clickhouse/clickhouse-server:latest
     container_name: clickhouse-keeper1
@@ -69,10 +74,10 @@ services:
       #   constraints: [node.labels.role == server]
       update_config:
         delay: 10s
-      # resources:
-      #   limits:
-      #     cpus: "0.50"
-      #     memory: 100M
+      resources:
+        limits:
+          cpus: "0.50"
+          memory: 1200M
     depends_on:
       - clickhouse-keeper1
       - clickhouse-keeper2
@@ -100,10 +105,10 @@ services:
       #   constraints: [node.labels.role == server]
       update_config:
         delay: 10s
-      # resources:
-      #   limits:
-      #     cpus: "0.50"
-      #     memory: 100M
+      resources:
+        limits:
+          cpus: "0.50"
+          memory: 1200M
     depends_on:
       - clickhouse-keeper1
       - clickhouse-keeper2
@@ -115,8 +120,10 @@ services:
 networks:
   clickhouse-server-network:
     driver: overlay
+    attachable: true
   clickhouse-keeper-network:
     driver: overlay
+    attachable: true
 
 volumes:
   clickhouse_data1: