Improve fallback logic on timeout #1
Description
If endpoint A times out once, we should use endpoint B for a while before trying A again, in order to avoid very slow RPC responses while endpoint A is temporarily degraded.
{{ message }}
If endpoint A times out once, we should use endpoint B for a while before trying A again, in order to avoid very slow RPC responses while endpoint A is temporarily degraded.