aws_templates

This repo contains template configuration files that we use to configure our AWS Elastic Beanstalk environment.

Overview

The list below describes features of the configuration with relevant files.

• Raid0
  • deploy.sh
• Low diskspace monitor
  • deploy.sh
• cron triggered log-rotation to AWS S3
  • environment.config
  • deploy.sh
  • check_cron.sh
  • logrotate-local
  • logrotate-tomcat-rc
  • logrotate.conf.elasticbeanstalk.httpd
  • logrotate.conf.local
  • logrotate.conf.tomcat.production
• AWS ELB configuration
  • environment.config
• System settings
  • limits.conf
  • sysctl.conf
• JMX remote monitoring
  • environment.config
  • catalina-jmx-remote.jar
• collectd monitoring
  • environment.config
  • collectd.conf
  • config_collectd.sh
  • collectd-librato.zip
• bypassing the default front-facing Apache web server and using Tomcat directly
  • httpd.conf
  • server.xml

tn_s3_file_uploader

We use the Ruby gem tn_s3_file_uploader to copy files from the EC2 instance to S3. You can fork the github project here and find usage documentation here.

Getting Started

Requirements

The provided files have been tested with

• ami-ed8e928, Amazon Linux AMI x86_64 PV S3 (Instance Store) 2014.03.1
• Beanstalk Stack Version Tomcat 7 Java 7 on 64bit Amazon Linux 2014.03 v1.0.4
• Beanstalk web server. Environment type load balancing, autoscaling
• Instances role that allows getting and putting objects to S3.
• Beanstalk option parameter PARAM2 is 'production'

Filling in the blanks

The provided files are missing working values for various settings. Make the following updates:

environment.config

• Replace JMX_USER with a username.
• Replace JMX_PASSWORD with a password.
• Replace JMX_ROLE with a access-level role. For monitoring, this can be "readonly".
• Replace JMX_PORT with a JMX listener port.
  • See Java docs for more details on remote JMX monitoring
• Replace S3_LOG_BUCKET with the bucket name of your log files destination.
• Replace AWS_ELB_BUCKET with the bucket name where your ELB Access logs will be dumped.

collectd.conf

• Replace JMX_USER with a username.
• Replace JMX_PASSWORD with a password.
• Replace JMX_PORT with a JMX listener port.
• Replace LIBRATO_USER with a Librato account email.
• Replace LIBRATO_EMAIL with a Librato account API Token. This token should have permission to write metrics.

logrotate.conf.tomcat.production

• Replace S3_LOG_BUCKET with the bucket name of your log files destination.

Adding it to your project

1. If your Java application is packaged as a WAR file, copy the .ebextensions to the same level as the WEB-INF directory. Otherwise, copy .ebextensions to the top-level directory of your project.

2. Copy your files to an S3 bucket. This bucket name should match the value of S3_FILES_BUCKET. Be sure that the instance role allows instances to getting and putting files from this bucket.

Environment stages and PARAM2

The environment stage is set with the Beanstalk option PARAM2. PARAM2 must be 'production' for this template configuration.