* Any contractual obligations on the part of the company which violate the laws of physics are the responsibility of sales and legal, not operations.
* All changes to the laws of physics must be submitted in RFC documents and subject to peer review before they will be considered valid for operational purposes.
* Except as defined in existing contracts, no service level agreement is offered in regard to predicting future events. All forecasts are to be evaluated in statistical context. For internal purposes, precognition is not subject to change by marketing or sales departments.
* In some cases it is not possible to give advance notice of unexpected outages, and it is not always possible to schedule hardware failures during maintenance windows.
* In most cases, causing change to past events will require upgrades to our time travel capabilities. Upgrades require capital expenditure, which must go through the standard approval process, with a business case clearly outlining expected costs and benefits (increased revenue or decreased expenditure). To save costs, our usual practice is to change past events as they are occurring instead of trying to correct them after the fact. When approval for past changes is denied, we will normally submit documentation at the time decisions are made as background for "I told you so" reviews later.
* 20th-century operating systems and hardware belong in museums, not on a production network.
* Windows sucks. If you are trying to run a web server using Windows, you're an idiot.
* Vendors lie. It's their job. When the vendor says, "This isn't our responsibility and we have no reason for outage for this incident as we did not resolve it," it means it is their responsibility and they did fix it, but they are so embarrassed by the cause that they won't admit anything could have happened that might have been their fault.
* Some vendors suck more than others. Usually the cheaper ones who are a lot cheaper suck a lot more.
* Actually, yes, we told you so. Please see attached.