Theres a good chance the index has been corrupted somehow. I'd suggest manually deleting the physical index files on the server, attempt a manual rebuild of the index, then ensure the optimize index scheduled task is enabled.
Technically speaking, if your task that runs to import data from another system into a custom table is executing, the index should be triggered assuming the cache key is broken since the data has changed. What I'd do is manually trigger an update to that index after the import succeeds in your import code.