T-SQL Tuesday is here once again and I need to be quick this time(I believe this is the second time I have to write a T-SQL Tuesday blog post in a 30 minute period). However, that does not mean that the message I am going to deliver to you is not important. It’s completely the opposite(or at least, I think so) and if I have to be honest, the post itself is not that technical. It’s more or less touching on the business side of supporting a service and should be read by IT managers and decision makers and not only by you, SQLFamily 🙂 . Let me tell you what I mean…
Less then a year ago I left HP to become a consultant and up until now, I was not engaged to create a monitoring solution of any kind(except for one recent nasty performance issue). That is something that makes me happy, if I have to be honest, because for the 3 years I spent as a Senior DBA for HP Enterprise Services, I have seen and done a lot for numerous Fortune 500s.
HP and all other major vendors that are in the “supporting services” business many times have their own monitoring solutions which makes a DBA’s life even easier because there is no need for creating or using Alerts, Jobs, Policy-Based Management and custom scripts(except if that monitoring software cannot give you what you need as a DBA). There’s a problem, though. Big companies(and not only) are sometimes slow(OK, not sometimes – very often!) and here is the moment when why I said this post is more for the IT decision makers, will clarify. When you, as a customer pay such company(huge amount of money) to support your SQL Servers and you are one of those few that really strives to be at the cutting edge of the technologies, please do yourself a favour and ask not just whether or not the team has the experience to deliver that service for you. Ask them to show you how are they going to monitor that same service. You want to see that in action. Think of that as a PoC on monitoring and here is a proof why you need to do that with just a simple example. Imagine this…
When SQL Server 2012 was released we saw the birth of this great technology called AlwaysOn Availability Groups(and many customers, for one reason or another, wanted it immediately). Your service provider, though – do they have a working monitoring solution for it or is it that the DBAs have to figure it out by themselves? Is that solution generating incidents(hopefully they have some ITIL framework implemented) that are giving the correct picture of the current status of your databases? Is that monitoring software also working just fine and not generating some absurd incident tickets just because it was not written well and adapted to that cool new piece of technology? To say it in simple words – is your service provider providing you an official guarantee that their monitoring system supports that new HA solution? I hope that you are probably understanding why I am asking those questions, right? Oh, one more – if they do not provide an official way(and the DBAs have to build their own solution) of supporting your service, do they plan to do it and how much time will it take to get it working – weeks, months, years? Recall from the beginning of the post – some companies are slow to adapt. This is where I want to leave you by warning you, IT managers, that you may be thinking that some service providers have everything set up in the most brutally perfect way, but that’s not always the case and I want you to know that and act accordingly.
Cathrine, thanks for the invitation and for this great topic! I am expecting quite some posts this time… 🙂
What do you think?