Internally check if the block processor needs to drop blocks.#4161
Internally check if the block processor needs to drop blocks.#4161clemahieu merged 1 commit intonanocurrency:developfrom
Conversation
thsfs
left a comment
There was a problem hiding this comment.
Noticed that the add_local function doesn't check for full (). Isn't it necessary there?
…than making it a requirement of the caller. Adds a stat counter when the block_processor drops a block as this is new behavior.
46ed8a1 to
cbca293
Compare
|
I'm not 100% sure about limiting the add_local, if the source of those blocks is RPC then overriding user action without any feedback might be unexpected. |
Agree on adding a feedback to the RPC for block processor is |
|
Looking through call sites, the add_local is only called by process_local_async, when the "process" RPC is called with "async". In our documentation we already mention observing when a block is inserted actually, via the websocket stream. Either we can document here or somewhere else that the block processing queue can internally drop and the stat to check for. |
Internally check if the block processor needs to drop blocks, rather than making it a requirement of the caller.
Adds a stat counter when the block_processor drops a block as this is new behavior.