When I moved proxies over to ansible, I made the max number of apache workers configurable, as we have some proxies that have more memory available than others. The default for the 'large' proxies is 900 apache workers. Our default nagios 'warn' is at 900 and critical is at 1000 for processes. We can and are easily hitting that limit on these larger proxies when they get a lot of requests + the normal pile of threads and such they have running all the time. So, I would like to change the nagios limits on those proxies. This will require applying the patch here and then running the proxy playbook. It should ideally result in less anoying nagios alerts where there is no action we actually need to take. +1s? kevin -- diff --git a/inventory/host_vars/proxy01.phx2.fedoraproject.org b/inventory/host_vars/proxy01.phx index 8f965b0..c9fbdcb 100644 --- a/inventory/host_vars/proxy01.phx2.fedoraproject.org +++ b/inventory/host_vars/proxy01.phx2.fedoraproject.org @@ -14,5 +14,5 @@ eth0_ip: 10.5.126.52 # This is consumed by the roles/fedora-web/main role sponsor: redhat -nrpe_procs_warn: 900 -nrpe_procs_crit: 1000 +nrpe_procs_warn: 1200 +nrpe_procs_crit: 1400 diff --git a/inventory/host_vars/proxy02.fedoraproject.org b/inventory/host_vars/proxy02.fedorapr index 4d4625e..21249f1 100644 --- a/inventory/host_vars/proxy02.fedoraproject.org +++ b/inventory/host_vars/proxy02.fedoraproject.org @@ -16,5 +16,5 @@ sponsor: internetx datacenter: internetx postfix_group: vpn -nrpe_procs_warn: 900 -nrpe_procs_crit: 1000 +nrpe_procs_warn: 1200 +nrpe_procs_crit: 1400 diff --git a/inventory/host_vars/proxy06.fedoraproject.org b/inventory/host_vars/proxy06.fedorapr index e807acb..055a56d 100644 --- a/inventory/host_vars/proxy06.fedoraproject.org +++ b/inventory/host_vars/proxy06.fedoraproject.org @@ -16,5 +16,5 @@ sponsor: osuosl datacenter: osuosl postfix_group: vpn -nrpe_procs_warn: 900 -nrpe_procs_crit: 1000 +nrpe_procs_warn: 1200 +nrpe_procs_crit: 1400 diff --git a/inventory/host_vars/proxy10.phx2.fedoraproject.org b/inventory/host_vars/proxy10.phx index 2cdcead..c16be55 100644 --- a/inventory/host_vars/proxy10.phx2.fedoraproject.org +++ b/inventory/host_vars/proxy10.phx2.fedoraproject.org @@ -13,3 +13,6 @@ eth0_ip: 10.5.126.51 # This is consumed by the roles/fedora-web/main role sponsor: redhat + +nrpe_procs_warn: 1200 +nrpe_procs_crit: 1400 _______________________________________________ infrastructure mailing list infrastructure(a)lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/infrastructure

When I moved proxies over to ansible, I made the max number of apache workers configurable, as we have some proxies that have more memory available than others. The default for the 'large' proxies is 900 apache workers. Our default nagios 'warn' is at 900 and critical is at 1000 for processes. We can and are easily hitting that limit on these larger proxies when they get a lot of requests + the normal pile of threads and such they have running all the time. So, I would like to change the nagios limits on those proxies. This will require applying the patch here and then running the proxy playbook. It should ideally result in less anoying nagios alerts where there is no action we actually need to take. +1s? kevin -- diff --git a/inventory/host_vars/proxy01.phx2.fedoraproject.org b/inventory/host_vars/proxy01.phx index 8f965b0..c9fbdcb 100644 --- a/inventory/host_vars/proxy01.phx2.fedoraproject.org +++ b/inventory/host_vars/proxy01.phx2.fedoraproject.org @@ -14,5 +14,5 @@ eth0_ip: 10.5.126.52 # This is consumed by the roles/fedora-web/main role sponsor: redhat -nrpe_procs_warn: 900 -nrpe_procs_crit: 1000 +nrpe_procs_warn: 1200 +nrpe_procs_crit: 1400 diff --git a/inventory/host_vars/proxy02.fedoraproject.org b/inventory/host_vars/proxy02.fedorapr index 4d4625e..21249f1 100644 --- a/inventory/host_vars/proxy02.fedoraproject.org +++ b/inventory/host_vars/proxy02.fedoraproject.org @@ -16,5 +16,5 @@ sponsor: internetx datacenter: internetx postfix_group: vpn -nrpe_procs_warn: 900 -nrpe_procs_crit: 1000 +nrpe_procs_warn: 1200 +nrpe_procs_crit: 1400 diff --git a/inventory/host_vars/proxy06.fedoraproject.org b/inventory/host_vars/proxy06.fedorapr index e807acb..055a56d 100644 --- a/inventory/host_vars/proxy06.fedoraproject.org +++ b/inventory/host_vars/proxy06.fedoraproject.org @@ -16,5 +16,5 @@ sponsor: osuosl datacenter: osuosl postfix_group: vpn -nrpe_procs_warn: 900 -nrpe_procs_crit: 1000 +nrpe_procs_warn: 1200 +nrpe_procs_crit: 1400 diff --git a/inventory/host_vars/proxy10.phx2.fedoraproject.org b/inventory/host_vars/proxy10.phx index 2cdcead..c16be55 100644 --- a/inventory/host_vars/proxy10.phx2.fedoraproject.org +++ b/inventory/host_vars/proxy10.phx2.fedoraproject.org @@ -13,3 +13,6 @@ eth0_ip: 10.5.126.51 # This is consumed by the roles/fedora-web/main role sponsor: redhat + +nrpe_procs_warn: 1200 +nrpe_procs_crit: 1400