CloudWatch Example Project
This project is a simple/stupid example of setting up and using CloudWatch Logging, Metrics, and Alerts.
Normally, you'd do this in an EC2 instance, but I thought it would be useful/convenient/interesting to automate and
demonstrate using Vagrant.
Prerequisites
-
Vagrant
The latest copy is the best bet.
-
AWS Credentials
The VM gets configured with AWS CLI credentials you provide on the command line. The credentials specified require appropriate CloudWatch permissions.
Getting Started
Once you've cloned this repo, just do a vagrant up.
OK, it's not necessarily quite that simple. You'll want your AWS credentials to be present in your
environment. If you use a credentials file, you'll want to do vagrant up, but provide the
variables along the way:
AWS_REGION="<region you like>" NOTIFICATION_EMAIL="your.email@yourdomain.com" DISCRIMINATOR="<some hopefully unique name>" AWS_ACCESS_KEY_ID="<your key>" AWS_SECRET_ACCESS_KEY="<your secret key>" vagrant up
(note: I've only tested in us-east-1, so it might be best to like it most :-\ )
The DISCRIMINATOR variable is used to separate your stuff by some useful name. Your first initial and last name is a good way to go.
Once started, Vagrant does the rest. Note that you will receive an SNS subscription confirmation email at the address you provided once the VM is configured. You must confirm the subscription in order to receive alert emails.
Experimenting
Once the VM is started, use vagrant ssh to get into the instance and look around. While you're in there,
be sure to do some stuff as as root, such as add new users (sudo useradd -m <someusername>) or just become root
with sudo su. All these activities are monitored by the CloudWatch Log Metrics put in place by the provisioning
script, so run a few cases through.
To see the effects of your activities as well as other activities performed by the included scripts, log into the AWS console and navigate to CloudWatch. Note the sub-sections of CloudWatch in the left-hand menu.
- Check CloudWatch Logs: type your discriminator into the filter box and hit enter. You should see a few Log Groups reflected. Click in to browse the logs.
- Check out Metrics: in the Metrics console click on PeChatMetrics. Investigate the metrics that have been
created by the script. Some metrics are separated by dimension (based on
DISCRIMINATOR) and some are not.
If you experimented as suggested above, the graphs should show some interesting activity (your own, and that of others that are running the code). - Experiment with Dashboards: try creating a Dashboard (Actions in the Graph Options tab of Metrics).