Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

add metrics for individual OVN components #138

Merged
merged 1 commit into from
Dec 10, 2024
Merged

add metrics for individual OVN components #138

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
33 changes: 33 additions & 0 deletions assets/ocp-performance/queries.libsonnet
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -164,6 +164,39 @@ local generateTimeSeriesQuery(query, legend) = [
query():
generateTimeSeriesQuery('topk(10, sum(container_memory_rss{pod=~"ovnkube-node-.*",namespace="openshift-ovn-kubernetes",container="ovn-controller"}) by (pod,node))', '{{pod}} - {{node}}'),
},
ovnKubeControllerCPU: {
query():
generateTimeSeriesQuery('topk(10, sum( irate(container_cpu_usage_seconds_total{pod=~"ovnkube-.*",namespace="openshift-ovn-kubernetes",container="ovnkube-controller"}[$interval])*100) by (pod,node) )', '{{pod}} - {{node}}'),
},
ovnKubeControllerMem: {
query():
generateTimeSeriesQuery('topk(10, sum(container_memory_rss{pod=~"ovnkube-node-.*",namespace="openshift-ovn-kubernetes",container="ovnkube-controller"}) by (pod,node))', '{{pod}} - {{node}}'),
},
topNbdbCPU: {
query():
generateTimeSeriesQuery('topk(10, sum( irate(container_cpu_usage_seconds_total{pod=~"ovnkube-.*",namespace="openshift-ovn-kubernetes",container="nbdb"}[$interval])*100) by (pod,node) )', '{{pod}} - {{node}}'),
},
topNbdbMem: {
query():
generateTimeSeriesQuery('topk(10, sum(container_memory_rss{pod=~"ovnkube-node-.*",namespace="openshift-ovn-kubernetes",container="nbdb"}) by (pod,node))', '{{pod}} - {{node}}'),
},
topNorthdCPU: {
query():
generateTimeSeriesQuery('topk(10, sum( irate(container_cpu_usage_seconds_total{pod=~"ovnkube-.*",namespace="openshift-ovn-kubernetes",container="northd"}[$interval])*100) by (pod,node) )', '{{pod}} - {{node}}'),
},
topNorthdMem: {
query():
generateTimeSeriesQuery('topk(10, sum(container_memory_rss{pod=~"ovnkube-node-.*",namespace="openshift-ovn-kubernetes",container="northd"}) by (pod,node))', '{{pod}} - {{node}}'),
},
topSbdbCPU: {
query():
generateTimeSeriesQuery('topk(10, sum( irate(container_cpu_usage_seconds_total{pod=~"ovnkube-.*",namespace="openshift-ovn-kubernetes",container="sbdb"}[$interval])*100) by (pod,node) )', '{{pod}} - {{node}}'),
},
topSbdbMem: {
query():
generateTimeSeriesQuery('topk(10, sum(container_memory_rss{pod=~"ovnkube-node-.*",namespace="openshift-ovn-kubernetes",container="sbdb"}) by (pod,node))', '{{pod}} - {{node}}'),
},

promReplCpuUsage: {
query():
generateTimeSeriesQuery('sum(irate(container_cpu_usage_seconds_total{pod=~"prometheus-k8s-0",namespace!="",name!="",container="prometheus"}[$interval])) by (pod,container) * 100', '{{pod}}')
Expand Down
26 changes: 17 additions & 9 deletions templates/General/ocp-performance.jsonnet
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -49,15 +49,23 @@ g.dashboard.new('Openshift Performance')
+ g.panel.row.withPanels([
panels.timeSeries.genericLegend('ovnkube-control-plane CPU Usage', 'percent', queries.ovnKubeControlPlaneCPU.query(), { x: 0, y: 1, w: 12, h: 8 }),
panels.timeSeries.genericLegend('ovnkube-control-plane Memory Usage', 'bytes', queries.ovnKubeControlPlaneMem.query(), { x: 12, y: 1, w: 12, h: 8 }),
panels.timeSeries.genericLegend('Top 10 ovn-controller CPU Usage', 'percent', queries.topOvnControllerCPU.query(), { x: 0, y: 9, w: 12, h: 8 }),
panels.timeSeries.genericLegend('Top 10 ovn-controller Memory Usage', 'bytes', queries.topOvnControllerMem.query(), { x: 12, y: 9, w: 12, h: 8 }),
panels.timeSeries.genericLegend('ovs-master CPU Usage', 'percent', queries.OVSCPU.query('$_master_node'), { x: 0, y: 17, w: 12, h: 8 }),
panels.timeSeries.genericLegend('ovs-master Memory Usage', 'bytes', queries.OVSMemory.query('$_master_node'), { x: 12, y: 17, w: 12, h: 8 }),
panels.timeSeries.genericLegend('ovs-worker CPU Usage', 'percent', queries.OVSCPU.query('$_worker_node'), { x: 0, y: 25, w: 12, h: 8 }),
panels.timeSeries.genericLegend('ovs-worker Memory Usage', 'bytes', queries.OVSMemory.query('$_worker_node'), { x: 12, y: 25, w: 12, h: 8 }),
panels.timeSeries.genericLegend('99% Pod Annotation Latency', 's', queries.ovnAnnotationLatency.query(), { x: 0, y: 33, w: 8, h: 8 }),
panels.timeSeries.genericLegend('99% CNI Request ADD Latency', 's', queries.ovnCNIAdd.query(), { x: 8, y: 41, w: 8, h: 8 }),
panels.timeSeries.genericLegend('99% CNI Request DEL Latency', 's', queries.ovnCNIDel.query(), { x: 16, y: 41, w: 8, h: 8 }),
panels.timeSeries.genericLegend('Top 10 ovnkube-controller CPU Usage', 'percent', queries.ovnKubeControllerCPU.query(), { x: 0, y: 10, w: 12, h: 8 }),
panels.timeSeries.genericLegend('Top 10 ovnkube-controller Memory Usage', 'bytes', queries.ovnKubeControllerMem.query(), { x: 12, y: 10, w: 12, h: 8 }),
panels.timeSeries.genericLegend('Top 10 ovn-controller CPU Usage', 'percent', queries.topOvnControllerCPU.query(), { x: 0, y: 18, w: 12, h: 8 }),
panels.timeSeries.genericLegend('Top 10 ovn-controller Memory Usage', 'bytes', queries.topOvnControllerMem.query(), { x: 12, y: 18, w: 12, h: 8 }),
panels.timeSeries.genericLegend('Top 10 nbdb CPU Usage', 'percent', queries.topNbdbCPU.query(), { x: 0, y: 26, w: 12, h: 8 }),
panels.timeSeries.genericLegend('Top 10 nbdb Memory Usage', 'bytes', queries.topNbdbMem.query(), { x: 12, y: 26, w: 12, h: 8 }),
panels.timeSeries.genericLegend('Top 10 northd CPU Usage', 'percent', queries.topNorthdCPU.query(), { x: 0, y: 34, w: 12, h: 8 }),
panels.timeSeries.genericLegend('Top 10 northd Memory Usage', 'bytes', queries.topNorthdMem.query(), { x: 12, y: 34, w: 12, h: 8 }),
panels.timeSeries.genericLegend('Top 10 sbdb CPU Usage', 'percent', queries.topSbdbCPU.query(), { x: 0, y: 42, w: 12, h: 8 }),
panels.timeSeries.genericLegend('Top 10 sbdb Memory Usage', 'bytes', queries.topSbdbMem.query(), { x: 12, y: 42, w: 12, h: 8 }),
panels.timeSeries.genericLegend('ovs-master CPU Usage', 'percent', queries.OVSCPU.query('$_master_node'), { x: 0, y: 50, w: 12, h: 8 }),
panels.timeSeries.genericLegend('ovs-master Memory Usage', 'bytes', queries.OVSMemory.query('$_master_node'), { x: 12, y: 50, w: 12, h: 8 }),
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

We should migrate "master" to "control-plane" holistically one day, but not a blocker here since you are only moving the lines around.

panels.timeSeries.genericLegend('ovs-worker CPU Usage', 'percent', queries.OVSCPU.query('$_worker_node'), { x: 0, y: 58, w: 12, h: 8 }),
panels.timeSeries.genericLegend('ovs-worker Memory Usage', 'bytes', queries.OVSMemory.query('$_worker_node'), { x: 12, y: 58, w: 12, h: 8 }),
panels.timeSeries.genericLegend('99% Pod Annotation Latency', 's', queries.ovnAnnotationLatency.query(), { x: 0, y: 66, w: 8, h: 8 }),
panels.timeSeries.genericLegend('99% CNI Request ADD Latency', 's', queries.ovnCNIAdd.query(), { x: 8, y: 66, w: 8, h: 8 }),
panels.timeSeries.genericLegend('99% CNI Request DEL Latency', 's', queries.ovnCNIDel.query(), { x: 16, y: 66, w: 8, h: 8 }),
]),
g.panel.row.new('Monitoring stack')
+ g.panel.row.withGridPos({ x: 0, y: 0, w: 24, h: 1 })
Expand Down