test_in_memory_trie_node_consistency is flaky: shard is currently congested and rejects new transactions

[jancionear]

Running test_in_memory_trie_node_consistency on the current master (23e22d5) causes a failure ~50% of the time.

The error message says something about the shard being congested:

$ cargo nextest run --cargo-profile dev-release --features statelessnet_protocol test_in_memory_trie_node_consistency
...
--- STDERR:              integration-tests tests::client::features::in_memory_tries::test_in_memory_trie_node_consistency ---
thread 'tests::client::features::in_memory_tries::test_in_memory_trie_node_consistency' panicked at integration-tests/src/tests/client/features/in_memory_tries.rs:289:62:
Invalid tx: Shard 1 is currently congested and rejects new transactions.
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace

   Canceling due to test failure
------------
     Summary [   9.062s] 1 test run: 0 passed, 1 failed, 2036 skipped
        FAIL [   6.961s] integration-tests tests::client::features::in_memory_tries::test_in_memory_trie_node_consistency
error: test run failed

Could be caused by #11419, /cc @jakmeier

[jancionear]

Fixed by #11508