site outage today (this time it was the heat)

[trainershack]

mine went down earlier due to leaves on the server...

on a different point, when i post a reply (using firefox) the site freezes on my computer and i have to reload the site. my reply is however submitted. any idea why it does this for me? when i use ie it's fine.

[slogger]

Hi Peter,

I am going to abide by the rules of the forum and not mention my URL or service until the magic 30 posts but them maybe some of you may like to check out my background...

Anyway, for what I can add:

1) Most IT service providers will quote an availability figure less than 100%. this means they are allowed downtime without constituting a service breach. My provider quotes 99.9% uptime, therefore the 0.1% downtime allowance constitutes 10 minutes (10080 minutes for a 24x7 service). If your provider is only quoting 99% uptime, he is allowed 100 minutes (just over 1 1/2 hrs per week).
The question "who do you trust" covers off my next point, how do they report on it and what penalties exist. My view of most web hosting services (and I apologies is any members of this forum offer a hosting service that does not follow these norms) is that they offer a service for your money but offer little in terms in terms of service level reporting and service management. If my host does not deliver against its 99.9% I have no service credit option and how do I prove that the downtime was not a problem with my ISP if they say that was the case

2) You state later on that the actual outage was as a result of a "failed change". This is totally unacceptable. If you really want to give them a hard time ask their customer service line for a copy of their problem managers "Major Incident Report" and a copy of the "Request For Change". All basic principles that a good IT service provider should have in house. With multiple servers all running similar OS level, no change should go in without a strong test program and backout plan.

3) With regard to air con and data centres, yep failures happen from individual fans, network cards and PSU in the server to main air handling units. Most good datacentres operate on a 1+x capacity so they have resiliance for failure. Mind you, that does not mitigate the possibility of a flood or a fire (both of which we have had in our computer room!)

PS have been in the forum for about a month now and am really enjoying it. Many thanks for a great place to meet, discuss and learn....

[TWD-Tony]

Whilst I agree with the majority of what you say Slogger, you make it sound like hosting companies use the 99% uptime advertisement as a get-out clause... Being the MD of a hosting company I can tell you that this isn't so...
The problem is... People expect CHEAP hosting - would anyone here pay for truely 100% uptime?
The amount of investment needed to create 100% redundancy is very large indeed, with multiple servers in different DC's running a round-robin DNS config... We are talking lots of money and this is just from the server hardware aspect alone... The network is another expense that needs massive investment to prevent a point of failure.
I think that for the majority of people (business and personal) the mid-priced 99% uptime packages are ideal.
Oh and Welcome to the forum

[slogger]

Hi (I assume its Tony),

I hope did not offend, as that was not my intention. You make a very truthful point and in my 4 years as a customer facing Service Delivery Manager, I have only had to use the "well we are allowed x minutes downtime" on a few occasions.

My view of service levels are that:
1) The give a clear expectation of offered service so customers can make an informed choice
2) Set a baseline for Service Improvement
3) Provide a basis for service reporting
4) Where needed give IT service providers (and Problem Managers!) the opportunity to keep a service down to get down to true root cause to prevent re-occurance (and I think that is where some of the angle of my thread was going)

With regard to your comment about cheap hosting, I must admit I am amazed at just how cheap "cheap is" now. I have only set my business up 8 weeks ago and with only a limited shopping around I was able to get 2 years unlimited hosting etc with a "quoted" 99.9% uptime for £35 per year. The whole set up costs inc all the registration with company house etc cost me under £250. Thats 1 days consultancy cost for the full 2 year outlay

I am interested in how as a service provider you report to your customers and how you deal with breaches of quoted availability?

I would also be interested in you casting you eye over my site to look at my "offering". As a supplier I would appreciate your take on it?

I can PM the URL to keep in the spirit of the rules

[peterjhale]

well we didn't take your advice and stayed with our current host (uh-hosting.co.uk) and as you can see from the last few days we have paid the price. Sorry about not being able to access the forum, server outages and just plain nail pulling slowness of everything.

As per before, uh are promising to sort these issues out but we shall see when all this resolved itself.