Hi Robert,
I think I've picked up on all of the questions here (there's a lot going on!) and have borrowed some awesomeness from my Tech Wizard (Boss) to fill in the exciting bits, so here goes:
I'll start with the easy one first... well actually, none of them are that hard
As a part of our ongoing SEO we check page load speed for our clients. A few days ago a client who has their site hosted by the same company was running very slow (about 8 seconds to load without cache). We ran every check we could and could not find a reason on our end. The client called the host and were told they needed to be on some other type of server (with the host) at a fee increase of roughly $10 per month.
OK, basically the answer to this one would be that your client's site was being throttled back by the host because it was using more bandwidth than was allowed under their existing plan. By moving them to the next plan (the extra $10 per month) the problem is resolved and the site speed returns to normal. Throttling it back gets the client to call... 8(
OK, 1 down and 2 to go...
About 4 months ago we realized we had a group of sites down thanks to monitoring alerts and checked it out. All were on the same IP address and the sites on the other IP address were still up and functioning well. When we contacted the support at first we were stonewalled, but eventually they said there was a problem and it was resolved within about 2 hours. Up until recently we had no problems.
and also
Yesterday, we noticed one group of sites on our server was down and, again, it was one IP address with about 8 sites on it.
OK, you know already that there can be up to 8 IPs on a box and at times something in the network will go bad. There are some variables here as to what is wrong. If you are on a Class C Network and one IP goes down then it means that the Switch or Router has gone bad (whether it is a Switch or a Router is determined by how the host has their hardware set up). If you are on a Class D Network and one IP goes down, then the problem is one of 3 things, the Card, the port, or the cable connecting the two, related to that IP.
The trick is that the person on the phone needs to realise what they are dealing with and escalate it to get the hardware issue resolved (A recent interaction with that particular host for one of our clients indicated to me that the realization part might be a little hit and miss, so good to have an understanding of what might be happening if it happens again)
Phew! Nearly there, last of all...
**On chat with support, they kept saying it was our ISP. (We speed tested on multiple computers and were 22MB down and 9MB up +/-2MB). We ran a trace on the IP address and it went through without a problem on three occassions over about ten minutes. After about 30 minutes the sites were back up. **
Here's the twist: we had a couple of people in the building who were on other ISP's try and the sites came up and loaded on their machines. Does anyone have any idea as to what the issue is?
OK this one is all about DNS caching. That particular host (the one that likes lady racing drivers) has a fail-over system in place. This means that if an IP goes down, the domains on that IP will automatically fail-over to another box.
So, if you have looked at those domains on your machine, it will be cached. When you go back to check the site you are still looking at the cached version. The other people in the building are coming to the domain fresh and through a different ISP, so they see those domains because they are back up on the new box.
When the host reps were telling you that it was your ISP, what they really meant was that it had failed-over to a new box and you were still seeing the cached DNS location.
OK, think I covered it all so....that's all Folks!
Have a great holiday weekend!
Sha