When a node is detected as unhealthy, we take the following actions:
Remove the node entirely from the routing.
Penalize its moving average by adding a specified value to the stored latency window. This ensures that any node exhibiting intermittent outages is appropriately penalized.
Previous approach was to simply remove the node from routing.
Extended the test_update_for_existing_node_succeeds().
Description
When a node is detected as unhealthy, we take the following actions:
Previous approach was to simply remove the node from routing.
Extended the
test_update_for_existing_node_succeeds()
.Checklist: