Add neutron ports alert rules

When neutron is under heavy load, routers can fail and leave ports
that are not binding. Alert in such condition can help operations

Change-Id: I36bddc1db00143d8773d762b8033aef0c52bbafc
This commit is contained in:
Gabriel Cocenza 2024-10-17 15:48:21 -03:00
parent 53d091b918
commit 94e7344c8e
No known key found for this signature in database

View File

@ -11,3 +11,27 @@ groups:
description: | description: |
The Neutron service '{{ $labels.service }}' is currently down on host {{ $labels.hostname }}. The Neutron service '{{ $labels.service }}' is currently down on host {{ $labels.hostname }}.
LABELS = {{ $labels }} LABELS = {{ $labels }}
- alert: NeutronPortsCritical
# router_gateway are ports that connects the router to an external network.
expr: openstack_neutron_port{admin_state_up="true", binding_vif_type="binding_failed", device_owner="network:router_gateway"} > 0
for: 5m
labels:
severity: critical
annotations:
summary: Neutron gateway ports binding failing.
description: |
Please check if neutron workers and/or server is overloaded
LABELS = {{ $labels }}
- alert: NeutronPortsWarning
# Other ports are very likely to be workload related, like creating new VMs
expr: openstack_neutron_port{admin_state_up="true", binding_vif_type="binding_failed", device_owner!="network:router_gateway"} > 0
for: 5m
labels:
severity: warning
annotations:
summary: Neutron ports binding failing.
description: |
Some ports are failing to bind
LABELS = {{ $labels }}