Overview
Important
This is a recommended release for op-batcher operators running with --data-availability-type=auto, and an optional one for everyone else. It contains corrected calldata pricing on L1 chains that have activated Glamsterdam, and the removal of the server-side RPC metrics.
Other changes
Data availability
- With
--data-availability-type=auto, the calldata-versus-blob cost comparison now prices calldata with the Glamsterdam gas rules once that fork is active on L1: a 15,000 gas base cost per transaction instead of 21,000, and a floor of 64 gas per calldata byte instead of 40. Previously the old model stayed in use after the fork, underestimating calldata and sometimes choosing it when blobs were cheaper. The fork is detected from the L1 block header, so no configuration is needed, and the numbers are unchanged on an L1 that has not activated it. TheUsing calldata channel configandUsing blob channel configlog lines gain anamsterdamfield showing which rules applied (#22996)
Metrics
- The server-side RPC metrics
op_batcher_default_rpc_server_*(requests, durations, responses, parameter and result sizes, notifications sent) are no longer exported. Dashboards and alerts that query them need updating; theop_batcher_default_rpc_client_*metrics are unchanged (#23114, #23055)
Miscellaneous
- When the batcher detects malformed transaction data it is about to submit, such as a calldata transaction carrying more than one frame, it now shuts down cleanly with a
Critical error detected, attempting batcher shut downerror log instead of exiting immediately with aCRITlog. Alerts matching the oldUnexpected number of frames in calldata txorUnexpected blob txdata with AltDA enabledlines need updating (#23054)
Full Changelog: op-batcher/v1.17.1...op-batcher/v1.17.2
🚢 Docker Image: