Comment 4 for bug 1918145

Hi,

thanks for the response.

After executing mysqlcheck -o -A -p (OPTIMIZE, ANALYZE) we did not manage to decrease the load and the number of rows examined.

- Yes we have [3].
- Here's a little of context for openstack network list:

Region running Stein release:
project: 1909
rbac rules: 22241
networks: 57
security groups: 2847

REPONSE TIMES:

AS ADMIN CONTEXT:
-----------------

real 0m8.596s
user 0m1.616s
sys 0m0.184s

AS USER CONTEXT:
----------------
real 0m16.632s
user 0m1.588s
sys 0m0.175s

We have successfully migrated to Ussuri one of our region, here's some figures:

project: 770
rbac rules: 29528
networks: 94
security groups: 1071

REPONSE TIMES:

AS ADMIN CONTEXT:
-----------------

real 0m5.163s
user 0m1.611s
sys 0m0.184s

AS USER CONTEXT:
----------------
real 0m6.398s
user 0m1.602s
sys 0m0.174s

I've noticed that somes route are taking a lot of time to respond. As we did not deploy the OSprofiler in production env, we do not know is they are stuck in python, or the DB, or even elsewhere:
- openstack security group list (56s)
- openstack network rbac list (18s)