Handle gracefully any unknown error that tries to fetch large block range

[desfero]

When RPC fails to fetch large block number request during eth_getLogs it does not always suggest smaller range.
Sometimes (as we seen on the scroll network) it tries to fetch large block range and timesout with an error similar to one below:
1:

{
    "jsonrpc": "2.0",
    "id": 42,
    "error": {
        "code": -32001,
        "message": "Unable to complete request at this time."
    }
}

2:

{
    "error": "Internal server error. Forwarder error: 1000."
}

Ridnexer needs to be smart and handle this by halving the block number range (we have this already for other errors) to not infinitely block indexing.