Merge bitcoin/bitcoin#34615: mempool: expose optimality of mempool to log / rpc

a9e59f7d95 rpc: add optimal result to getmempoolinfo (Greg Sanders)
a3fb3dd55c mempool: log if we detect a non-optimal mempool (Greg Sanders)

Pull request description:

  Post-SFL #34023 I don't think we expect the mempool to be unordered for long periods of time. If we consider it likely to be a serious regression in production, it would be useful to expose the fact  that the mempool is not known to be optimal.

  1. do a MEMPOOL log after any `DoWork()` returns false, meaning non-optimal
  2. expose it via getmempoolinfo, by calling `DoWork(0)`, which does nothing but return known-optimality

  I'm not wedded to either approach, I just think something is better than nothing for the next release.

ACKs for top commit:
  ajtowns:
    ACK a9e59f7d95
  ismaelsadeeq:
    reACK a9e59f7d95 [c89b93b95..a9e59f7d95](c89b93b958..a9e59f7d95) fixed typo, added more logging for block/reorg additions to mempool, and fixed brittle test case.
  sedited:
    ACK a9e59f7d95
  sipa:
    ACK a9e59f7d95

Tree-SHA512: 1560ad21cc1606df7279c102f35f61d4555c0ac920f02208b2a6eb89b14d7e22befb6d7f510a00a9074c2f9931f32e9af86bcea3a8dd9a1d947b0398c84666dd

test/functional/mempool_cluster.py

@@ -306,6 +306,9 @@ class MempoolClusterTest(BitcoinTestFramework):
 
         assert_equal(node.getrawmempool(), [])
 
+        # Key should exist and be trivially optimal
+        assert node.getmempoolinfo()["optimal"]
+
         # Not in-mempool
         not_mempool_tx = self.wallet.create_self_transfer()
         assert_raises_rpc_error(-5, "Transaction not in mempool", node.getmempoolcluster, not_mempool_tx["txid"])
@@ -367,6 +370,9 @@ class MempoolClusterTest(BitcoinTestFramework):
         chunkfee = first_chunk_tx["fee"] + second_chunk_tx["fee"] + third_chunk_tx["fee"]
         assert_equal(first_chunk_info, {'clusterweight': first_chunkweight + second_chunkweight + third_chunkweight, 'txcount': 3, 'chunks': [{'chunkfee': first_chunk_tx["fee"], 'chunkweight': first_chunkweight, 'txs': [first_chunk_tx["txid"]]}, {'chunkfee': second_chunk_tx["fee"], 'chunkweight': second_chunkweight, 'txs': [second_chunk_tx["txid"]]}, {'chunkfee': third_chunk_tx["fee"], 'chunkweight': third_chunkweight, 'txs': [third_chunk_tx["txid"]]}]})
 
+        # We expect known optimality directly after txn submission
+        assert node.getmempoolinfo()["optimal"]
+
         # If we prioritise the last transaction it can join the second transaction's chunk.
         node.prioritisetransaction(third_chunk_tx["txid"], 0, int(third_chunk_tx["fee"]*COIN) + 1)
         first_chunk_info = node.getmempoolcluster(first_chunk_tx["txid"])