comment out the "kernel.sched_migration_cost_ns"

This tests did not show a significant difference before and after turning a process scheduler - https://gitlab.com/postgres-ai/postgresql-consulting/tests-and-benchmarks/-/issues/16

To really reduce the number of context switches:
- try to minimize the number of idle connections to the database;
- use the connections pool in transaction mode.

in addition, this parameter is missing in ubuntu 22.04

Fixed:

failed: [172.31.22.219] (item={'name': 'kernel.sched_migration_cost_ns', 'value': '5000000'}) => {"ansible_loop_var": "item", "changed": false, "item": {"name": "kernel.sched_migration_cost_ns", "value": "5000000"}, "msg": "Failed to reload sysctl: vm.overcommit_memory = 2\nvm.swappiness = 1\nvm.min_free_kbytes = 102400\nvm.dirty_expire_centisecs = 1000\nvm.dirty_background_bytes = 67108864\nvm.dirty_bytes = 536870912\nvm.zone_reclaim_mode = 0\nkernel.numa_balancing = 0\nsysctl: cannot stat /proc/sys/kernel/sched_migration_cost_ns: No such file or directory\n"}

vars/system.yml

@@ -56,7 +56,6 @@ sysctl_conf:
 #    - {name: "vm.nr_hugepages", value: "9510"}  # example "9510"=18GB
     - {name: "vm.zone_reclaim_mode", value: "0"}
     - {name: "kernel.numa_balancing", value: "0"}
-    - {name: "kernel.sched_migration_cost_ns", value: "5000000"}
     - {name: "kernel.sched_autogroup_enabled", value: "0"}
     - {name: "net.ipv4.ip_nonlocal_bind", value: "1"}
     - {name: "net.ipv4.ip_forward", value: "1"}
@@ -66,6 +65,7 @@ sysctl_conf:
     - {name: "net.core.somaxconn", value: "65535"}
     - {name: "net.ipv4.tcp_tw_reuse", value: "1"}
 #    - {name: "net.netfilter.nf_conntrack_max", value: "1048576"}
+#    - {name: "kernel.sched_migration_cost_ns", value: "5000000"}
 #    - {name: "", value: ""}
   balancers:
     - {name: "net.ipv4.ip_nonlocal_bind", value: "1"}