Hi Robert,
That makes sense. One other oddity I've noticed is that the database CPU
actually goes down when I spin up more nodes, for the same number of
users/queries per second.
--Josh
On Wed, Dec 13, 2023 at 4:12 PM robert engels wrote:
> It is contending with sql.rows.awaitDone().
>
> My
It is contending with sql.rows.awaitDone().
My guess is that when you close rows it has to wait for db acknowledgement -
that this is a false positive based on wall time and not cpu time - that the
mutex is waiting for the cleanup operation to complete.
> On Dec 13, 2023, at 1:25 PM, Joshua