New AWS Console Feature - Improved Access to CloudWatch Alarms

We've added some features to the AWS Management Console to make it easier for you to create, view, and manage your Amazon CloudWatch Alarms. Here's a visual overview.

Creating Alarms
You can now create a new CloudWatch alarm from the Monitoring tab of the selected EC2 instance or EBS volume using the Create Alarm button:

Let's say I want to know when the Network Out traffic for my JeffServer instance exceeds 1.5 Megabytes within a 5 minute period (this instance hosts personal blogs for me and several members of my family, along with some other random web sites, none of which see a whole lot of traffic). I chose this number after inspecting the detailed graph for this metric on this instance:

A click of the Create Alarm button takes me to the new Create Alarm dialog. I can choose the metric and the time interval, and I can also choose the notification method:

I can choose to send notifications to an existing SNS topic, create a new topic, or to a list of email addresses. If I choose the latter option, CloudWatch will automatically create an SNS topic with a suitable name and subscribe the email addresses to the topic. The dialog displays the alarm threshold using a red line superimposed on the actual metrics data:

The Monitoring tab now displays a summary of the alarms for the selected instance (highlighting added):

Clicking on the summary displays a list of alarms:

We hope that you enjoy (and make use of) this handy new feature!

-- Jeff;

Comments

[Frankmunz]

Hi Jeff,

What is missing is the last step: AWS console should provide a way to shut down an instance based on an alarm setting.
I know it can be done with autoscaling, but a more direct way, accessible from console would be desirable. Many users are afraid of excessive costs that could occur, e.g. when sb accesses a page on CloudWatch/S3 repeatingly to produce costs eg. on a competitor site.

cheers,

Frank

--
Middleware and Cloud Computing
http://www.amazon.com/dp/0980798000/

[Christopher]

What would be nice is a global monitor alert when AMZN is going to do hypervisor maintenance. An email with something like:
The following instances have maintenance scheduled:

instance-number (name) at Wed Dec 21 between 9pm an 11pm (PST)
other-instance (name2) at Thus Dec 22 between 6pm and 9pm (PST)
etc.

Not everyone logs into the console every day ;)