Investigate if table scans on the configuration context could cause performance issues

At the moment some of the queries JIRA runs on the configurationcontext database table perform table scans. This does cause problems as the number of records in this table is small. The numebr of records does depend on the number of custom fields that are set up and how these are configured (i.e. it depends on the number of contexts, project/issue type configurations, each custom field has).

We should investigate if the table scans could become a problem in the future, and if we can/should put more indexes in place on the configurationcontext table.