Hi,
I collected the new Circuit build expiry summary counters over a few hours of runtime and got these totals:
prefix_hops_0 0
prefix_hops_1 2
prefix_hops_2 120
post_prefix 0
open_channel 122
other_channel 0
connection_failed 0
late_firsthop 2
completed 284
abandoned 60
excluded 0
adaptive 1
What caught my attention is the very strong concentration at prefix_hops_2: 120 observations, compared with only 2 at prefix_hops_1 and none at prefix_hops_0 or post_prefix.
For reference:
prefix_hops_2 / completed = 120 / 284 = 42.3%
I understand that this ratio is not necessarily a circuit failure rate, since these counters may represent different sets of observations.
My main question is simply: is such a large imbalance toward prefix_hops_2 expected/normal?
And does prefix_hops_2 effectively mean that the circuit successfully completed two hops but expired while trying to extend to the third hop?
There are no connection_failed or other_channel events in this sample.
Thanks