Node lagging

[saniksin]

Lately, I’ve noticed that my node can fall behind the main chain by 100–200 blocks even under low server load, then later catch up without any errors in the logs. This happens both on testnet and mainnet, across different servers.

When emulating transaction loads (80–100 tx within 1–2 seconds), the node may also start dropping and lagging behind the network, even though CPU, disk (writes), and system load can remain low.

Have you investigated this issue or tried optimizing performance under medium–high load?

I also noticed that sometimes a Starknet feeder gateway endpoint doesn’t respond. Could you add support for multiple endpoints (as reserve/fallback endpoints), so the node can automatically switch in case of an error or timeout?

https://feeder.alpha-sepolia.starknet.io/feeder_gateway/get_state_update?blockNumber=1810150&includeBlock=true

https://feeder.alpha-mainnet.starknet.io/feeder_gateway/get_state_update?blockNumber=1810150&includeBlock=true

[saniksin]

https://t.me/c/2182307908/3691 - Starknet feeder gateway endpoint doesn’t respond.

https://t.me/c/2182307908/3734 - node got stuck

I’ve encountered this issue many times.

[kkovaacs]

@saniksin Are you still experiencing these issues now that Starknet mainnet has (hopefully) stabilized after the 0.14.0 upgrade?

[saniksin]

@saniksin Are you still experiencing these issues now that Starknet mainnet has (hopefully) stabilized after the 0.14.0 upgrade?

I've noticed that during high network load, when blocks are produced every 1-3 seconds, my node sometimes lags behind the chain. However, when the network slows down, it gradually catches up. I'm not sure if this is related to hardware or a software issue, but my server itself doesn’t show high load.