Prior to this PR, we were very aggressive in fetching the block count. During startup, as fast as we could. After startup, ~3 block count fetchers were started, when only one was needed. That in turn, lead to "work queue depth exceeded".
After this PR, we only start one connection timer per chain. One for mainchain, another for sidechain. Also remove the super-aggressive startup-fetcher. In sum, we are much nicer to the node.
Prior to this PR, we were very aggressive in fetching the block count. During startup, as fast as we could. After startup, ~3 block count fetchers were started, when only one was needed. That in turn, lead to "work queue depth exceeded".
After this PR, we only start one connection timer per chain. One for mainchain, another for sidechain. Also remove the super-aggressive startup-fetcher. In sum, we are much nicer to the node.