Starting at around midnite last night, Telstra crashed their primary, secondary and management servers.
Any customers using telstra for DNS were outaged for a significant time today more than 8 hours.
Here is a discussion about the problems – http://forums.whirlpool.net.au/forum-replies-archive.cfm/1492786.html
Think about it – you have a single box sitting running your server.
It sits there for 18 months, NO downtime at all. Wow, I bet your boss is impressed.
The Drama
Now the battery on the raid card, is swollen (because of heat and age), and disrupts the RAID card. BANG, your server is down.
Here is a common SLA from rackspace .
The SLA kicks in AFTER they “identify the cause of the problem”. So you have to lodge a ticket – perhaps this takes you 30 minutes to recognise the problem, find your credentials, and lodge a ticket. The hosting provider has to get a techo to your server, reboot it a few times, atleast an hour is passed. Then perhaps they take another 30 minutes to find its the RAID card.
They now have 1 hour to repair it in the SLA “Hardware replacement is guarantied to be complete within one hour of problem identification.”
So there is 3 hours, your customers and your boss are screaming, no SLA refunds yet, and no solution.
The SLA says “We will credit your account 5% of the monthly fee per additional hour of downtime, up to 100% of your monthly fee for the affected server(s).”
Now the Rub
So you might get like $120 refund then they replace the raid card from onsite spares in the fourth hour. It gets much worse, if they have to wait for a hardware vendor to bring attend on site to replace the raid card.
What was the cost of 4 hours outage to your company in terms of cash, lost sales etc, and what is the damage to your brand.
Who is going to tell your boss that you have a $120 refund coming on next month’s hosting bill. Not me !
Are you now and going to ring up and yell at the agency that recommended the hosting provider. I wouldn’t want to take that call either.
What’s worse, is the hosting provider has met their obligations under the contract, and you are stuck for another 6-18 months.
A better way
Brandprotector does not use physical servers. We have automatic failover and clustering. We use vmware virtual servers, that are just as fast, but can instantly recover from failed server parts.
Add to that our brand guarantees, all the way up to $1million. How much harder for your uptime do you think we work ? $120 penalty vs $1million penalty ?
Given our penalty is similar to yours, we have a teaming approach. We understand it might be your job, your clients job, and our job on the line.
When hosting buyers see 99.95% uptime guarantee, they think “great, that’s what I need”.
In reality, the contract says there is a penalty for not reaching that level. Its generally 10% of that months contract.
So if you have revenues of say $50 million p.a., and pay for hosting for say $2,000 month. If the hosting company does outage your service, they will rebate you $200. Do you think that will help in any way to alleviate your woes ? unlikely.
The contract often says, even on an 8 hour outage, that you get 50% of the current month refunded.
The contract even says, they are not in breach, and you are expected to continue the contract for the period.
So at the end of the day, the SLA and contract protects your provider, not you.
At brand protector, the contract protects all of us, including the agency and your client.
