We need more intel about what's going on with blocks / transactions / operations / mempool
That would help with getting a better view on situations like this:
2021-11-15T14:45:20.680239 block_producer.cpp:178 apply_pending_transa ] Postponed 206 transactions due to block size limit
2021-11-15T14:45:20.735940 p2p_plugin.cpp:704 broadcast_block ] Broadcasting block #59199619
2021-11-15T14:45:20.737028 witness_plugin.cpp:344 block_production_loo ] Generated block #59199619 with timestamp 2021-11-15T14:45:21 at time 2021-11-15T14:45:21
Currently in "live sync" we see informations like this:
2021-11-16T12:56:06.013 p2p_plugin.cpp:187 handle_block ] Got 200 transactions on block 59226170 by gtg -- Block Time Offset: 13 ms
Would be great to have a better info on:
- number of transactions
- number of operations
- size of block in bytes
- with some extra switch even more data on mempool
also --set-benchmark-interval
could be extended to provide statistical information on data processed within given interval:
- avg block size
- max block size
- number of transactions
- number of operations
Some of that info we have because of periodic RocksDB reports (if account_history_rocksdb
plugin is enabled), some are logged during hive sync
, still having that available at seed nodes could help with ad-hoc State-of-Hive analysis.