I'm running nextjs in a kubernetes cluster and when my pod starts it has this huge memory spike up to 700mb. Any recommendations how to avoid that?