What is your strategy for handling sudden spikes in log volume that threaten to overwhelm your logging cluster?
