swarm-monitoring

Docker Swarm monitoring and instrumentation with Prometheus and ELK stack

Prerequisites

Docker Swarm cluster:

• Docker Engine min version 17.09
• 3 manager nodes
• 3 worker nodes
• VPC with urestricted access between nodes

Docker hosts vm map:

sudo sysctl -w vm.max_map_count=262144
sudo echo 'vm.max_map_count=262144' >> /etc/sysctl.conf

Disable swap in /etc/fstab and run:

sudo swapoff -a

Grand group access to local dir bind mount:

mkdir /esdatadir
chmod g+rwx /esdatadir
chgrp 1000 /esdatadir

Overview

ELK Services:

• Elastic Search Data (3 nodes)
• Elastic Search Router
• Elastic Search Exporter
• Elastic Search Curator
• Kibana
• Logstash (global)
• Logsprout (global)

Deploy ELK:

./elk-evo-up.sh <SWARM-MANAGER-IP>

Deploy the reverse proxy on the jump host:

./caddy-up.sh