detailed (comparison) view: The cache misses column should be removed, invocations should be renamed #529

Since we don't record cache misses for perf.rlo, it does not make sense to show the number of cache misses (it will always be the same as the "# of invocations" right next to it).

When recording cache misses and cache hits, then invocations is hits + misses. When not recording cache hits, then invocations is misleading because it only counts the cache misses.

On zulip we settled on "full executions" as the new title for the "invocations" column. This subsumes "query provider invocations" and regular (thus uncached) function invocations.