SQL Probe 7.0.44 1 2 3 4 5 6 7 8 9

Monitoring servers is normally a full time task for one or more DBAs.  In fact, the DBA job is getting more and more fragmented as environments get larger.  We now have operational DBAs, development DBAs, replication DBAs, lead DBAs, data modelers, etc.  One person used to perform all of these functions and it is still this way in smaller environments.  In environments where you have 20+ SQL Servers, this is just not possible.  Each of these functions requires specialized skills and also specialized tools.

For those who have managed large environments, you know that one of the biggest hurdles is getting all of the important information for your servers into a central location to work with.  Having to connect to several dozen or several hundred servers to get information can consume the majority of your time.  In most cases, you simply want to ensure that the server is up, accessible, and that none of the important services has crashed.  You may have written a small app or a proc that runs at a specified interval to simply connect to all of your servers.  SQL Probe brings all of this functionality down into a single interface.  It will periodically connect to each of your registered servers and poll it for system, database, and process information.

The main screen shown below gives you all of the basic information you need to ensure your servers are up and accessible.  The icons across the top represent each of the major processes.  They are from left to right: SQL Server, SQL Agent/Executive, databases, SQL Mail, DTC, Full Text Search, and replication.  They have attempted to keep the interface the same between versions of SQL Server.  This has benefits and drawbacks.  I understand why they have done this, but I don't agree with it.  For example, on a 6.5 server there will be an icon showing that full text search is not installed.  While this is true, it is also not possible to do with a 6.5 server.  The benefits are that the interface looks exactly the same between versions.  The drawback is that it can get very confusing.  The current version and service pack of each of the SQL Servers that you have registered will be displayed below the icons.  The remainder of the server view gives you process and network information.  The CPU percentage is very useful to allow you to get a quick snapshot of how heavily loaded a server is.  If a user complains about slow response time when the server is at 100% CPU and when it is at less than 10% CPU are two completely different situations.  In the first case, it is likely that nothing is wrong on the server and that processes are just queuing up for processor resources.  In the other case, you probably have a blocked process.  The processes show as a percentage of the maximum.  I'm unclear about what this would mean on a 7.0 server since connections are dynamically allocated and there really isn't a min or max for processes.  Since they are dynamically allocated, I would assume that my processes should always show as 100% on a 7.0 server.  This is not happening since I have 50 processes running against a server and the processes are still showing as 0%.  Total I/O and total network are relatively useless when monitoring.  SQL Probe is meant to be started and left running.  The total I/O and total network are simply an accumulation of counters across the entire period.  This could be an hour, day, week, month, year.  Looking at the total I/O for an entire month gives me no information as to what is going on now or what occurred in the past.