The lack of decent tools to manage your databases on an enterprise level is just staggering. It's hard to believe that nobody can cross the finish line on delivering a monitoring solution.
One of my most recent disappointments is Quest Foglight. I've been dealing with Foglight for going on 2yrs now I guess, and it's just getting worse and worse. For starters, its capabilities were grossly exaggerated by the sales force and the pre-sales team. Now, I'm stuck with a solution that is proving itself to be fairly unstable, and hard to manage. On paper it looked very attractive. It monitored DB and OS stats, sent alerts, provided realtime graphs that were configurable, etc. It had a lot to offer.
Unfortunately, we have come to discover that each repository cannot handle very many monitored hosts. I'm told now that a single Foglight repository could handle no more than 200 hosts, when I was originally told that it could handle 3x that many. I need to clarify this though... that's 200 default hosts. Once you start adding others that number goes down significantly. A default host is like the NT agent. For SQL Server, Oracle, LogReader, Web Server, etc, decrease that number.
We have come across specific performance counters that don't even collect data, an often times confusing security model, problems with customizing the graphs, long troubleshooting sessions with no results.
The bugs are plentiful as well. There are many times when the Foglight agent drives the CPU of the server to 98% and we have to kill the service to return to our production schedule. Foglight only alerts us to system outages part of the time, and sees no problems with many downed systems. Agents just go offline for no apparent reason, and though we can restart them, they go back down in just a few minutes... or a few days, we never really know. Oddly enough, even the help files that come with it manage to corrupt themselves every couple months, and we have to replace them.
The database they chose to implement this with is Faircom. This was an architectural blunder that belongs on the same echelon as with running with scissors. They do have an Oracle version now, but from what Quest just told me, we won't really see any improvements in the new version.
Users can't manage their own security either. You have to assign them security accounts, and keep up with the passwords yourself. Why do you have to do this? Well, if you place a user into a security group, and need to either move him to another group, or add him to another group, you have to supply his password. I don't even understand that decision. You don't see that anywhere else. AD admins aren't expected to know the passwords of all the users in the enterprise, nor are SQL or Oracle admins. In fact, NOBODY else is has this limitation. It would be nice to be able to tie the login in with LDAP, but again, it's not going to happen for a while.
The web interface they provide for remote administration is just a joke. It doesnt honor the security of the security groups you defined in the java console, so if there's a server you don't want a tech to see, it's highly likely that he will be able to log onto the web and see it that way... way to go Quest!
Not only are these problems plentiful, they are also not going away anytime soon. I've had many discussion with Quest about their roadmap, and it looks long and slow. I don't see anything major happening even by this time next year. There are many more bugs and lack of functionality than I mentioned here though.
Quest is a huge company, and it's starting to show. Their customer support has all but completely gone away. I get such incredibly bad technicians every time I call in, and even the high-level customer service managers are only interested in politics. I have yet to have any of them actually solve any of my problems.
In short, Quest talks a good game, but really falls short when it comes to delivering. Do your research before you buy Foglight because it is not scalable at all, and if you have more than say 50 servers, you will regret your decision.
Later this week I'll talk to you about the problems with Quest Spotlight. Talk about failure to deliver... more on that later.