Skip to main content

For Home Stakers

For one validator or a small fleet on your own hardware. The questions this page helps with: which client combination runs reliably on a NUC or mini-PC, which upgrades to apply, what to actually monitor.

A note on hardware scale

Our published numbers come from a bare-metal fleet (Epyc 9654P, NVMe, dedicated NICs), heavier than typical home-staker hardware. Relative comparisons (which client is faster, which uses less memory) carry over. Absolute numbers are an upper bound for modest hardware.

Where to start

  • Read the blog: most posts answer questions a home staker actually has, like "which client should I run?", "what does this upgrade mean for me?", "is this cloud setup OK?".
  • How we measure: a short read on methodology, so you can judge how to apply our numbers to your setup.
  • Free plan: 1 Grafana user, 7-day delay, 90-day retention. Useful for browsing dashboards and comparing your node to ours.

Posts a home staker should bookmark

Operational rules of thumb from our measurements

  • Open inbound TCP+UDP 30303 if you're on a cloud or behind a strict NAT. Without it your EC peer count is roughly 30% of what it should be, and your CC won't tell you.
  • Besu's default --max-peers=25 is the lowest of any EC. Consider raising to 50.
  • Heavy CC choice matters more than light CC choice for resource cost. Lighthouse, Lodestar, and Grandine sit on the higher peer end. Prysm and Nimbus are more conservative.
  • Don't trust beacon_head_slot from your CC to know if your EC is synced. Four of six ECs allow their CC to report "in sync" while the EC is still hours away. Check EC logs.

Stay updated

Subscribe via RSS. New posts roughly weekly. The FAQ covers common operator questions.

For questions or content requests: contact@stereumlabs.com.