11: Monitoring
CloudWatch Overview
The analogy used for why we monitor: a car driving towards a cliff that we want to monitor if the brake works.
CloudWatch is a monitoring and observability platform that was esigned to give us insight into our AWS architecture. It allows us to monitor multiple levels of our applications and identify potential issues.
- System metrics: metrics you get out of the box. The more managed the service, the more you get. E.g. "my CPU is too high", "my memory usage is too much".
- Application Metrics: by installing the CloudWatch agent, you can get info from inside your EC2 instances.
- Alarms: alert you when something goes wrong.
There are two types of metrics:
- Default: Things out-of-the-box.
- Custom: When you need to install the agent.
When creating an alarm:
- Alarms for custom metrics can go down to check every 10 seconds, but default metrics minimum is 1 minute.
- To be notified, you can use SNS to let you know.
The tips:
- Anytime monitoring comes up, think CloudWatch.
- There are no defualt alarms. You must create them yourself.
- Default vs. Custom: AWS cannot see past the hypervisor level for EC2 instances.
- The more managed the service, the more you get out of the box.
- The stardard metric interval is 5 minutes. Detailed is 1 minute that you need to opt into.
Application Monitoring with CloudWatch Logs
What do we do with all our logs for our instances? CloudWatch Logs.
Part of the CloudWatch suite, CloudWatch Logs is a service that allows you to monitor, store and access log files from a variety of different sources.
It gives you the ability to query your logs to look for potential issues or data that is relevant to you.
Terms
- Log Event: The record of what happened. It contains a timestamp and the data.
- Log Stream: A collection of log events from the same source create a log streams. Think of one continuous set of logs from a single instance.
- Log Group: A collection of log streams. For example, you'd group all of your Apache web server logs across hosts together.
CloudWatch Logs Features
- Filter Patterns: You can look for specific terms in your logs. E.g. searching for 400 errors.
- CloudWatch Logs Insights: Use SQL-like commands to search through all logs.
- Alarms: Alert you when something goes wrong.
Installing the CloudWatch Agent
On an AMI that uses yum:
$ sudo yum install amazon-cloudwatch-agent -y # Prompts with questions, including a question about what log files to monitor.
Exam tips:
- Generally you will favour CloudWatch Logs unless the exam asks for a real-time solution.
- Alarms can be used to alert if the filter patterns are found.
- CloudWatch agent must be installed and configured. It's not automatic.
- If the question mentions SQL, think CloudWatch Logs Insights.