[1.0.3 -> main] P2P: Schedule a retry of enqueue_sync_block when throttled

[heifner]

When a connection is throttled it will not attempt to enqueue another block unless enqueue_sync_block is called. This will happen on any message sent to the connection. However, that can be a long time if the node has to wait for a heart-beat timeout before some message is sent to the peer.

Instead, schedule a try-again call to enqueue_sync_block in 100ms.

This fixes the current issue with p2p_sync_throttle_test which could not make progress because it repeately received a sync wait timeout.

Merges release/1.0 into main including #899

Partially resolves #882, See #897

[spoonincode]

needs to be based on gh_882_main unless you want two PRs merged in one upmerge

[ericpassmore]

Note:start
category: System Stability
component: P2P
summary: Schedule a retry of enqueue_sync_block when throttled.
Note:end