When DataWolf starts, it pages through the executions to find tasks that didn't finish when it was shut down. However, with a lot of executions, this is very slow. We should consider ways to query just for unfinished executions and not page through everything or handle this in a separate background thread so the engine isn't locked until this finishes