 |
| "Services", "EC2", "Instances", select the client's EC2 instance. Go to "Monitoring" tab, click "Create Alarm". |
 |
| Tick "Take the action". Choose "Recover this instance". It will then auto-choose the correct type of check (which is "Status Check Failed (System)"). Click "Create Alarm". |
 |
| Alarm created. Click "Close". |
 |
| At the beginning, the alarm may display a "insufficient data" warning. It can be safely ignored as the alarm has only just started collecting data. |
 |
| After a few minutes, it should be "OK". |
 |
| For the curious, more details about the CloudWatch alarm that we have just created. |
There is no way to emulate a "Status Check Failed (System)". So we have to trust Amazon that the alarm can recover the EC2 instance. This method of making an EC2 instance highly-available is taken from here.
Next, backup and restore...
Part 6 - IAM Role
Single EC2 Instance - Table of Contents
