It seems you've outlined a lot of the major points of network management.  Your goal should be to find tools which you find acceptable in terms of usability and usefulness.  Generally speaking, network management solutions should be able to automatically discover the network with minimal input from the user.  They should be able to graphically display the network, and provide clear, real-time information about fault conditions.  How fault conditions are determined should be based on leading industry practices, but should also be flexible so that you can build your own fault conditions, and eliminate those that are not applicable to your environment.  Remember, you want to be able to manage by exception.  You want your NMS to be able to filter out "noise" and provide with you with relevant, concise fault information.  If you can get fault management to a point where false positives have mostly been eliminated, then certainly, a solution which feeds those faults directly into a trouble ticketing solution is a good thing.  However, you will want more than just a case which says, "Device X is down."  Your solution should be able to track and provide details about what was happening to that device prior to the fault (e.g. inventory changes, config changes, syslog messages, performance data, etc.).

On the configuration and element management side, you will want general inventory data.  This includes things like hardware information (e.g. device type, port counts, module counts, memory stats, flash info, etc.).  You will also want asset tracking information (e.g. part numbers, serial numbers, etc.) as that will feed directly into your trouble ticketing system to facilitate opening support requests, and tracking inventory.  A good configuration management application will be able to track configuration changes in real-time.  It will show you what changes were made on the devices, when they were made, and by whom.  You should be able to easily spot configuration problems, and quickly revert configurations as needed.  Deploying new configuration changes to a large number of devices is also a must.  And any changes made should feed into some kind of accounting (or change control) component so you can pre-review changes before they are deployed.

Performance can be quite involved.  Generally, most people expect a performance management solution to be able to show them utilization (both network and device) at any point in their network.  But performance should also be tied into fault.  That is, you should be able to establish a performance baseline with thresholds so that you can be notified when performance begins to suffer (but BEFORE things become so bad that users start to complain).  On top of general performance, you should look at what kind of information will be traversing your network.  If you are a voice/video user, then you will want to make sure your NMS solution can measure quality of voice, jitter, latency, etc.  If you are using QoS tagging, you will want something that can illustrate the amount of traffic being classified in each QoS level; and show you the net effect of QoS on the network. Again, this should tie into fault management.  You want to be notified if QoV, jitter, etc. starts to become an issue.

There are quite a few NMS solutions which can make use of the Netflow and IP SLA features of IOS to provide comprehensive performance metrics.  Such solutions would be able to give a more complete picture of not only what is happening on the network, but also how your network is handling certain traffic.

And as I said in the first paragraph, usability should also be a concern.  I am not saying to sacrifice features for usability, but from experience, I can say that a more use-friendly, easy to understand application can save lots of time (and money).  This is especially true when you are trying to find root cause, or track down high-impact problems (i.e. those problems where you start to lose X number of dollars per minute).  So make sure the solution has meaningful, and easy-to-read reports, graphs, and displays.  You want the information (especially real-time information) to be useful to your NOC operators, but you will probably also want clear, concise reports for management to assess the quality of the network.