Multiple indexing issues detected on Optimism

[Uxio0]

We detected multiple indexing issues:

• A Safe that migrates from L1 to L2 will always trigger and error and then call reindex_master_copies when processing, as the status is always corrupt before the nonce the Safe had when it was migrated. Transactions are missing as the events were not emitted, and service tries to find the missing events forever.
• InternalTxDecoded.objects.safes_pending_to_be_processed() is really slow, taking more than 50 seconds in network like Optimism. That prevents a quick processing of Safes.
• Banned Safes don't mark InternalTxDecoded as processed, making the index and the table grow but never consumed.