Improve the retry logic to detect excessive retries

[felipet]

If the chosen value for the block_range is too high for the RPC, the request seems to timeout, which triggers a retry in the indexer's logic. This is very common when using third party RPCs.

If paying for usage of the RPC, this behaviour consumes tokens meaningless. The retry logic should detect whether a request was sent multiple times without success to stop the indexer, and inform the user that the chosen block_range is perhaps inadequate for that RPC.