How should we evaluate the performance of an DBA?

Can someone tell us a good way to measure the DBAs performance?

I am thinking of the problem of how to evaluate the performance of an DBA these days. As every one knows the DBA do not create the profit for the enterprise but they can protect the profit from the loss. It seems the more the enterprise loss the more the DBA create. Sounds paradox but it is true in many situations. Then the problem of how to evaluate the performance of an DBA is coming. In fact it is very hard to judge performance for any intelligent work not only the DBA jobs. In recent year some one published a lot of papers regarding how to evaluate the software developer work. But very few about the DBA’s. For the product DBAs is it they are very busy meaning they are working harder. Or the opposite is correct? On the contrary if an DBA do many things proactively and deal with the issue before it really happen. Not every one can find what he have done let alone the busy manager. Then how to let the manager know what he has done? Sometimes the problem solution is very complex but when asked it is simpler. No one have enough time know it unless they want to replace your work. Say there is a DBA he do a lot things proactively and Databases are always running smoothly but the developers even the manager do not notice his existence sometimes. All of them think the database should be running like this. The database should not throw any error. Maybe because the database product is good, maybe it is because our data volume is not big or something. They do not want to put up with any tiny error. They will scold the DBA even the issue is caused by storage or networks or application or terrible code. The DBAs are always blamed until it is proven they are innocent. Take another example the DBA do things reactively they see the errors happen sometimes but take any action until they are informed. The database refuses to continue work. Of course The DBA can handle it and make it resume working ASAP. Then many people think such kind of DBA is good DBA because they can handle any situation. Their performance is very good in the enterprise. As a good DBA we should spend 40% of our time on the issue resolved including the rootcauase finding, do the search and daily routine work. The other 60% of time we should use to do proactive things including predict the issue, analyze the trend, set up the monitor and do the heath check for the database and accumulating and research the new feature of the database product. Anyway the DBA work also needs a lot of communication skill. We need to talk with the customers to talk with our manager and even our co worker including the other DBAs and also the developers sometimes. I do not think the one who is busy with the issue researching; Performance tuning is a good DBA. We should set aside enough time to strengthen our weapon. To lean enough knowledge and skill to qualified us to the future challenge. Only in this way can we handle the battle under our control. Some departs set many plans at the beginning of a year and to see how many plans the DBA make them come true. They hope this can measure the DBAs performance. They also use the extra working hours in the past or how many incidents they handle in the past to measure their performance.

Does it sound reasonalbe?