support confluent connectors for apache install

[christopherkeller]

confluent supports ES sinks with a library package. we should install this for apache mode by default.

https://www.confluent.io/hub/confluentinc/kafka-connect-elasticsearch

[christopherkeller]

Connector uploaded into S3 bucket, but need IAM work on each EC2 instance to make it seamless. listed are the manual steps after uploading into S3.

• sudo yum install -y unzip
• sudo unzip /tmp/confluentinc-kafka-connect-elasticsearch-5.3.1.zip -d /usr/share/java
• sudo chown -R kafka:kafka /usr/share/java/confluentinc-kafka-connect-elasticsearch-5.3.1
• sudo systemctl restart apache-kafka-connect (this seems unecessary as the connectors picked up the connectors instantly)

verify connector is running

• curl http://`ifconfig eth1 | grep -e 'inet\s' | awk '{ print $2 }'`:8083/connector-plugins

On ansible host provision connectors as usual

AWS_PROFILE=slower ec2_region=us-east-2 ./deploy connect $platform_root/ucsd/inventory/hostsfile-us-east-2-demo.none demo aws us-east-2 development centos $platform_root/ucsd $key_path