Open samlaf opened 4 days ago
preprod is unstable, it is more difficult to identify.
testnet now is 3 mins
Indeed bowen just showed me testnet has been updated to 3 min batches: https://github.com/Layr-Labs/eigenda-devops/blob/master/charts/batcher/values/eigenda-testnet/us-east-1/holesky/values.yaml#L15
what does "difficult to identify" mean? Do we have other tricks to accelerate this test then? perhaps not waiting for finality?
difficult to identify the root cause when there is error. Moreover, oncall has responsibility to fix testnet, whereas it is possible for preprod to dangle for a while
Finality is taking a larger chunk of time now, 3mins vs 12mins. But I think it makes sense to test for finality.
Update: seems like e2e tests still took 14 mins without waiting for finality (action from #160)... so there's prob something else causing them to be slow.
testnet has 10 min batching intervals, which I think (?) is why the holesky-e2e tests currently take ~20 mins to run. preprod cluster has 1 min batching, which should reduce the ci time for that test.