Linux Administration Cookbook

How to do it...

Once the issue is recreated, you can get down to the meat of the work:

1. To begin, you need to know which boxes are likely to be connected to the problem, either directly, or by one degree of separation. For example, if the company website is down, I would immediately log on to whichever system hosted the website, while trying to find information on any systems that supplied it, such as messaging brokers or databases.
2. If you don't know what the topology of the deployment looks like off the top of your head, and you can't find a document in your internal systems that details the layout, it can be a good idea to engage others in the process of troubleshooting. 

You might not feel entirely comfortable about waking someone else ...