Electra: No aggregated attestation for slot

[twoeths]

Describe the bug

this was found by @nflaig

this could be due to #7546

Expected behavior

• at least, we should have 1 attestation from ourself. Or explain the reason
• but first, need to separate this metric for api vs gossip lodestar_attestation_pool_insert_outcome_total

Steps to reproduce

No response

Additional context

No response

Operating system

Linux

Lodestar version or commit hash

v1.27.1

[twoeths]

some api attestations may come late and it's after the cutoff time of the attestation pool. In that case the attestation pool should accept it because the getAggregate() call only comes after that and api attestation has higher priority

[nflaig]

this still happens on our holesky-prod fleet, see logs from last 6 hours, we can use this query to check once we have deployed a fix

also we should probably add a metric to better keep track of get aggregate cache misses

lodestar/packages/beacon-node/src/chain/opPools/attestationPool.ts

Lines 171 to 173 in 7ce4939

	if (!aggregate) {
	// TODO: Add metric for missing aggregates
	return null;

[nflaig]

some api attestations may come late and it's after the cutoff time of the attestation pool. In that case the attestation pool should accept it because the getAggregate() call only comes after that and api attestation has higher priority

while I agree that this is a good change I am not really convinced that it is the root cause that we have aggregate cache misses

just looking at the logs of holesky-prod-0 from last 30 days it seems the issue has started after the electra fork