Monitoring RDS instances

- [Instructor] Monitoring your database is a key aspect of maintaining the availability and performance of your application. AWS offers a variety of metrics to monitor your RDS instances. This includes network throughput, input/output for read and write operations, and client connections. To start with, it is important to establish a baseline for normal RDS performance in your environment. This should be done by measuring performance at different times and under different load conditions. As RDS is being monitored, consider storing historical monitoring data. This store data will help you build a baseline to compare against. When performance falls outside of the established baseline, it might be needed to change the instance class of your DB instance or the number of DB instances and read replicas to improve database availability for your workload. AWS recommends involving your network and database administrators and comparing against your baseline when investigating issues with CPU, RAM, network traffic, and database connections. The Amazon RDS console can be used to monitor basic metrics and investigate event logs. From the RDS management console, click the database name and then click the Monitoring tab. This allows you to monitor basic metrics such as CPU Utilization, Free Storage Space, Write IOPS, and Read IOPS. AWS also recommends using RDS events and database log files to watch RDS and report when something is wrong. RDS uses simple notification service, also known as SNS, to provide notification when an RDS event occurs. RDS groups these events into categories that you can subscribe to in order to be notified. Some of the RDS event categories include availability, backup, configuration change, creation and deletion, failure, failover, and low storage. The full list of RDS events can be found on the AWS RDS documentation. Subscribing to an RDS event notification is easy. From the RDS console, navigate to Event Subscriptions on the left navigation pane and then click Create Event Subscription. Start by providing a name for the subscription. Then select where you'd like to send the notifications to. We can select an SNS topic, an email topic, or an SMS topic. I'll select an SNS topic that I created earlier. Next, we'll select the source type. Different event types have different sources. For this demonstration, I'll set the source type to instances. Then select if you'd like to be notified from all objects in that resource type and across all event categories. Finally, I'll click the Create button. Please refer to the AWS RDS documentation to know the right event source. In addition to event notifications, RDS database logs can be viewed and downloaded. To view your database log files, click Databases on the left navigation pane. Then click on the name of your database, and navigate to the Logs and Events tab. Scroll down to the Logs section, select the log that you'd like to view, and click the View button. I'll click Close for now. The database instance logs can also be published to Amazon CloudWatch. CloudWatch provides a number of metrics across multiple dimensions to monitor your RDS instances. This will allow you to perform real-time analysis of log data in addition to setting up alarms and custom dashboards for monitoring. The full list of dimensions and metrics can be found on the AWS documentation. In addition to these, the AWS Trusted Advisor can also be used to check for idle DB connections, security group access risk, and optimizations.