You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
276 lines
10 KiB
276 lines
10 KiB
local g = import 'grafana-builder/grafana.libsonnet';
|
|
|
|
{
|
|
grafanaDashboards+:: {
|
|
'node-cluster-rsrc-use.json':
|
|
local legendLink = '%s/dashboard/file/node-rsrc-use.json' % $._config.grafana_prefix;
|
|
|
|
g.dashboard('USE Method / Cluster')
|
|
.addRow(
|
|
g.row('CPU')
|
|
.addPanel(
|
|
g.panel('CPU Utilisation') +
|
|
g.queryPanel(|||
|
|
(
|
|
instance:node_cpu_utilisation:rate1m{%(nodeExporterSelector)s}
|
|
*
|
|
instance:node_num_cpu:sum{%(nodeExporterSelector)s}
|
|
)
|
|
/ scalar(sum(instance:node_num_cpu:sum{%(nodeExporterSelector)s}))
|
|
||| % $._config, '{{instance}}', legendLink) +
|
|
g.stack +
|
|
{ yaxes: g.yaxes({ format: 'percentunit', max: 1 }) },
|
|
)
|
|
.addPanel(
|
|
// TODO: Is this a useful panel? At least there should be some explanation how load
|
|
// average relates to the "CPU saturation" in the title.
|
|
g.panel('CPU Saturation (load1 per CPU)') +
|
|
g.queryPanel(|||
|
|
instance:node_load1_per_cpu:ratio{%(nodeExporterSelector)s}
|
|
/ scalar(count(instance:node_load1_per_cpu:ratio{%(nodeExporterSelector)s}))
|
|
||| % $._config, '{{instance}}', legendLink) +
|
|
g.stack +
|
|
// TODO: Does `max: 1` make sense? The stack can go over 1 in high-load scenarios.
|
|
{ yaxes: g.yaxes({ format: 'percentunit', max: 1 }) },
|
|
)
|
|
)
|
|
.addRow(
|
|
g.row('Memory')
|
|
.addPanel(
|
|
g.panel('Memory Utilisation') +
|
|
g.queryPanel(|||
|
|
instance:node_memory_utilisation:ratio{%(nodeExporterSelector)s}
|
|
/ scalar(count(instance:node_memory_utilisation:ratio{%(nodeExporterSelector)s}))
|
|
||| % $._config, '{{instance}}', legendLink) +
|
|
g.stack +
|
|
{ yaxes: g.yaxes({ format: 'percentunit', max: 1 }) },
|
|
)
|
|
.addPanel(
|
|
g.panel('Memory Saturation (Major Page Faults)') +
|
|
g.queryPanel('instance:node_vmstat_pgmajfault:rate1m{%(nodeExporterSelector)s}' % $._config, '{{instance}}', legendLink) +
|
|
g.stack +
|
|
{ yaxes: g.yaxes('rps') },
|
|
)
|
|
)
|
|
.addRow(
|
|
g.row('Network')
|
|
.addPanel(
|
|
g.panel('Net Utilisation (Bytes Receive/Transmit)') +
|
|
g.queryPanel(
|
|
[
|
|
'instance:node_network_receive_bytes_excluding_lo:rate1m{%(nodeExporterSelector)s}' % $._config,
|
|
'instance:node_network_transmit_bytes_excluding_lo:rate1m{%(nodeExporterSelector)s}' % $._config,
|
|
],
|
|
['{{instance}} Receive', '{{instance}} Transmit'],
|
|
legendLink,
|
|
) +
|
|
g.stack +
|
|
{
|
|
yaxes: g.yaxes({ format: 'Bps', min: null }),
|
|
seriesOverrides: [
|
|
{
|
|
alias: '/ Receive/',
|
|
stack: 'A',
|
|
},
|
|
{
|
|
alias: '/ Transmit/',
|
|
stack: 'B',
|
|
transform: 'negative-Y',
|
|
},
|
|
],
|
|
},
|
|
)
|
|
.addPanel(
|
|
g.panel('Net Saturation (Drops Receive/Transmit)') +
|
|
g.queryPanel(
|
|
[
|
|
'instance:node_network_receive_drop_excluding_lo:rate1m{%(nodeExporterSelector)s}' % $._config,
|
|
'instance:node_network_transmit_drop_excluding_lo:rate1m{%(nodeExporterSelector)s}' % $._config,
|
|
],
|
|
['{{instance}} Receive', '{{instance}} Transmit'],
|
|
legendLink,
|
|
) +
|
|
g.stack +
|
|
{
|
|
yaxes: g.yaxes({ format: 'rps', min: null }),
|
|
seriesOverrides: [
|
|
{
|
|
alias: '/ Receive/',
|
|
stack: 'A',
|
|
},
|
|
{
|
|
alias: '/ Transmit/',
|
|
stack: 'B',
|
|
transform: 'negative-Y',
|
|
},
|
|
],
|
|
},
|
|
)
|
|
)
|
|
.addRow(
|
|
g.row('Disk IO')
|
|
.addPanel(
|
|
g.panel('Disk IO Utilisation') +
|
|
// Full utilisation would be all disks on each node spending an average of
|
|
// 1 second per second doing I/O, normalize by metric cardinality for stacked charts.
|
|
// TODO: Does the partition by device make sense? Using the most utilized device per
|
|
// instance might make more sense.
|
|
g.queryPanel(|||
|
|
instance_device:node_disk_io_time_seconds:rate1m{%(nodeExporterSelector)s}
|
|
/ scalar(count(instance_device:node_disk_io_time_seconds:rate1m{%(nodeExporterSelector)s}))
|
|
||| % $._config, '{{instance}} {{device}}', legendLink) +
|
|
g.stack +
|
|
{ yaxes: g.yaxes({ format: 'percentunit', max: 1 }) },
|
|
)
|
|
.addPanel(
|
|
g.panel('Disk IO Saturation') +
|
|
g.queryPanel(|||
|
|
instance_device:node_disk_io_time_weighted_seconds:rate1m{%(nodeExporterSelector)s}
|
|
/ scalar(count(instance_device:node_disk_io_time_weighted_seconds:rate1m{%(nodeExporterSelector)s}))
|
|
||| % $._config, '{{instance}} {{device}}', legendLink) +
|
|
g.stack +
|
|
{ yaxes: g.yaxes({ format: 'percentunit', max: 1 }) },
|
|
)
|
|
)
|
|
.addRow(
|
|
g.row('Disk Space')
|
|
.addPanel(
|
|
g.panel('Disk Space Utilisation') +
|
|
g.queryPanel(|||
|
|
sum without (device) (
|
|
max without (fstype, mountpoint) (
|
|
node_filesystem_size_bytes{%(nodeExporterSelector)s, %(fsSelector)s} - node_filesystem_avail_bytes{%(nodeExporterSelector)s, %(fsSelector)s}
|
|
)
|
|
)
|
|
/ scalar(sum(max without (fstype, mountpoint) (node_filesystem_size_bytes{%(nodeExporterSelector)s, %(fsSelector)s})))
|
|
||| % $._config, '{{instance}}', legendLink) +
|
|
g.stack +
|
|
{ yaxes: g.yaxes({ format: 'percentunit', max: 1 }) },
|
|
),
|
|
),
|
|
|
|
'node-rsrc-use.json':
|
|
g.dashboard('USE Method / Node')
|
|
.addTemplate('instance', 'up{%(nodeExporterSelector)s}' % $._config, 'instance')
|
|
.addRow(
|
|
g.row('CPU')
|
|
.addPanel(
|
|
g.panel('CPU Utilisation') +
|
|
g.queryPanel('instance:node_cpu_utilisation:rate1m{%(nodeExporterSelector)s, instance="$instance"}' % $._config, 'Utilisation') +
|
|
{
|
|
yaxes: g.yaxes('percentunit'),
|
|
legend+: { show: false },
|
|
},
|
|
)
|
|
.addPanel(
|
|
// TODO: Is this a useful panel? At least there should be some explanation how load
|
|
// average relates to the "CPU saturation" in the title.
|
|
g.panel('CPU Saturation (Load1 per CPU)') +
|
|
g.queryPanel('instance:node_load1_per_cpu:ratio{%(nodeExporterSelector)s, instance="$instance"}' % $._config, 'Saturation') +
|
|
{
|
|
yaxes: g.yaxes('percentunit'),
|
|
legend+: { show: false },
|
|
},
|
|
)
|
|
)
|
|
.addRow(
|
|
g.row('Memory')
|
|
.addPanel(
|
|
g.panel('Memory Utilisation') +
|
|
g.queryPanel('instance:node_memory_utilisation:ratio{%(nodeExporterSelector)s, %(nodeExporterSelector)s, instance="$instance"}' % $._config, 'Memory') +
|
|
{ yaxes: g.yaxes('percentunit') },
|
|
)
|
|
.addPanel(
|
|
g.panel('Memory Saturation (Major Page Faults)') +
|
|
g.queryPanel('instance:node_vmstat_pgmajfault:rate1m{%(nodeExporterSelector)s, instance="$instance"}' % $._config, 'Major page faults') +
|
|
{
|
|
yaxes: g.yaxes('short'),
|
|
legend+: { show: false },
|
|
},
|
|
)
|
|
)
|
|
.addRow(
|
|
g.row('Net')
|
|
.addPanel(
|
|
g.panel('Net Utilisation (Bytes Receive/Transmit)') +
|
|
g.queryPanel(
|
|
[
|
|
'instance:node_network_receive_bytes_excluding_lo:rate1m{%(nodeExporterSelector)s, instance="$instance"}' % $._config,
|
|
'instance:node_network_transmit_bytes_excluding_lo:rate1m{%(nodeExporterSelector)s, instance="$instance"}' % $._config,
|
|
],
|
|
['Receive', 'Transmit'],
|
|
) +
|
|
{
|
|
yaxes: g.yaxes({ format: 'Bps', min: null }),
|
|
seriesOverrides: [
|
|
{
|
|
alias: '/Receive/',
|
|
stack: 'A',
|
|
},
|
|
{
|
|
alias: '/Transmit/',
|
|
stack: 'B',
|
|
transform: 'negative-Y',
|
|
},
|
|
],
|
|
},
|
|
)
|
|
.addPanel(
|
|
g.panel('Net Saturation (Drops Receive/Transmit)') +
|
|
g.queryPanel(
|
|
[
|
|
'instance:node_network_receive_drop_excluding_lo:rate1m{%(nodeExporterSelector)s, instance="$instance"}' % $._config,
|
|
'instance:node_network_transmit_drop_excluding_lo:rate1m{%(nodeExporterSelector)s, instance="$instance"}' % $._config,
|
|
],
|
|
['Receive drops', 'Transmit drops'],
|
|
) +
|
|
{
|
|
yaxes: g.yaxes({ format: 'rps', min: null }),
|
|
seriesOverrides: [
|
|
{
|
|
alias: '/Receive/',
|
|
stack: 'A',
|
|
},
|
|
{
|
|
alias: '/Transmit/',
|
|
stack: 'B',
|
|
transform: 'negative-Y',
|
|
},
|
|
],
|
|
},
|
|
)
|
|
)
|
|
.addRow(
|
|
g.row('Disk IO')
|
|
.addPanel(
|
|
g.panel('Disk IO Utilisation') +
|
|
g.queryPanel('instance_device:node_disk_io_time_seconds:rate1m{%(nodeExporterSelector)s, instance="$instance"}' % $._config, '{{device}}') +
|
|
{ yaxes: g.yaxes('percentunit') },
|
|
)
|
|
.addPanel(
|
|
g.panel('Disk IO Saturation') +
|
|
g.queryPanel('instance_device:node_disk_io_time_weighted_seconds:rate1m{%(nodeExporterSelector)s, instance="$instance"}' % $._config, '{{device}}') +
|
|
{ yaxes: g.yaxes('percentunit') },
|
|
)
|
|
)
|
|
.addRow(
|
|
g.row('Disk Space')
|
|
.addPanel(
|
|
g.panel('Disk Space Utilisation') +
|
|
g.queryPanel(|||
|
|
1 -
|
|
(
|
|
max without (mountpoint, fstype) (node_filesystem_avail_bytes{%(nodeExporterSelector)s, %(fsSelector)s, instance="$instance"})
|
|
/
|
|
max without (mountpoint, fstype) (node_filesystem_size_bytes{%(nodeExporterSelector)s, %(fsSelector)s, instance="$instance"})
|
|
)
|
|
||| % $._config, '{{device}}') +
|
|
{
|
|
yaxes: g.yaxes('percentunit'),
|
|
legend+: { show: false },
|
|
},
|
|
),
|
|
),
|
|
},
|
|
}
|