Thanks Mike. Adding that index manually has definitely dropped the query
time. I was seeing the entry in the log every 2 minutes on average, even
with an idle MAAS.
What I am doing is logging postgres queries that take longer than a
threshold, currently 50ms. I will file bugs on those, then drop the
threshold to 40ms, and repeat.
My MAAS is small (20 nodes, 50 devices, 4 controllers) but I think as
part of the next cycle we should repeat this exercise using data from a
much larger MAAS.
Thanks Mike. Adding that index manually has definitely dropped the query
time. I was seeing the entry in the log every 2 minutes on average, even
with an idle MAAS.
What I am doing is logging postgres queries that take longer than a
threshold, currently 50ms. I will file bugs on those, then drop the
threshold to 40ms, and repeat.
My MAAS is small (20 nodes, 50 devices, 4 controllers) but I think as
part of the next cycle we should repeat this exercise using data from a
much larger MAAS.
Mark