We might expect a DBA (Apps Admin / SysAdmin / Ops Manager) to be able to answer questions like :
a. What has been the Server CPU Utilisation / IO Wait Time 1 month / 3 months / 6 months ago ?
b. What was the size of the (specific) Table / Schema / Database 1 month / 3 months / 6 months / 1 year / 2 years ago ?
c. Which application programs and/or SQL statements have been the cause of most complaints of performance (throughput / response time) or waits / "hangs" (really meaning 'enqueues') from the application / users / management and what has been done to monitor and/or improve them ? (statistics are necessary to provide evidence).
d. What errors occur in the database and whether patches have been identified / requested in the past 3 months / 6 months ?
e. What has been the volume of transactions in the past 1 month / 3 months / 6 months / 1 year ? As a corollary, what have been user volumes ? (Unfortunately, far too often, some manager might think that there is a high correlation between user volumes and transaction volumes / performance / throughput.)
Much of the evidence / metrics can be statistics. Some of it can be (unformatted) notes and documents and/or email trails.