def test_row_show_title():
row = G.Row().to_json_data()
assert row['title'] == 'New row'
assert not row['showTitle']
row = G.Row(title='My title').to_json_data()
assert row['title'] == 'My title'
assert row['showTitle']
row = G.Row(title='My title', showTitle=False).to_json_data()
assert row['title'] == 'My title'
assert not row['showTitle']
def run(self):
templateList = [
G.Template(default="",
dataSource="default",
name="serverid",
label="ServerID",
query="label_values(serverid)")
]
dashboard = G.Dashboard(title=self.options.title,
templating=G.Templating(list=templateList))
# Simple table processing - could be enhanced to use GridPos etc.
for metric in metrics:
if 'section' in metric:
dashboard.rows.append(
G.Row(title=metric['section'], showTitle=True))
continue
if 'row' in metric:
dashboard.rows.append(G.Row(title='', showTitle=False))
continue
graph = G.Graph(title=metric['title'],
dataSource='default',
maxDataPoints=1000,
legend=G.Legend(show=True,
alignAsTable=True,
min=True,
max=True,
avg=True,
current=True,
total=True,
sort='max',
sortDesc=True),
yAxes=G.single_y_axis())
ref_id = 'A'
for texp in metric['expr']:
graph.targets.append(G.Target(expr=texp, refId=ref_id))
ref_id = chr(ord(ref_id) + 1)
dashboard.rows[-1].panels.append(graph)
# Auto-number panels - returns new dashboard
dashboard = dashboard.auto_panel_ids()
s = io.StringIO()
write_dashboard(dashboard, s)
print("""{
"dashboard": %s
}
""" % s.getvalue())
def test_auto_refids_preserves_provided_ids():
"""
auto_ref_ids() provides refIds for all targets without refIds already
set.
"""
dashboard = G.Dashboard(
title="Test dashboard",
rows=[
G.Row(panels=[
G.Graph(
title="CPU Usage by Namespace (rate[5m])",
targets=[
G.Target(
expr='whatever #Q',
legendFormat='{{namespace}}',
),
G.Target(
expr='hidden whatever',
legendFormat='{{namespace}}',
refId='Q',
),
G.Target(
expr='another target'
),
],
).auto_ref_ids()
]),
],
)
assert dashboard.rows[0].panels[0].targets[0].refId == 'A'
assert dashboard.rows[0].panels[0].targets[1].refId == 'Q'
assert dashboard.rows[0].panels[0].targets[2].refId == 'B'
def test_auto_refids():
"""auto_ref_ids() provides refIds for all targets without refIds already set."""
dashboard = G.Dashboard(
title="Test dashboard",
rows=[
G.Row(panels=[
G.Graph(
title="CPU Usage by Namespace (rate[5m])",
dataSource="My data source",
targets=[
G.Target(
expr='whatever #Q',
legendFormat='{{namespace}}',
),
G.Target(
expr='hidden whatever',
legendFormat='{{namespace}}',
refId='Q',
hide=True
),
G.Target(
expr='another target'
),
],
yAxes=[
G.YAxis(format=G.SHORT_FORMAT, label="CPU seconds"),
G.YAxis(format=G.SHORT_FORMAT),
],
).auto_ref_ids()
]),
],
)
assert dashboard.rows[0].panels[0].targets[0].refId == 'A'
assert dashboard.rows[0].panels[0].targets[1].refId == 'Q'
assert dashboard.rows[0].panels[0].targets[2].refId == 'B'
def test_auto_id():
"""auto_panel_ids() provides IDs for all panels without IDs already set."""
dashboard = G.Dashboard(
title="Test dashboard",
rows=[
G.Row(panels=[
G.Graph(
title="CPU Usage by Namespace (rate[5m])",
dataSource="My data source",
targets=[
G.Target(
expr='whatever',
legendFormat='{{namespace}}',
refId='A',
),
],
yAxes=[
G.YAxis(format=G.SHORT_FORMAT, label="CPU seconds"),
G.YAxis(format=G.SHORT_FORMAT),
],
)
]),
],
).auto_panel_ids()
assert dashboard.rows[0].panels[0].id == 1
def Row(
panels, height=None, title='Dashboard Row', showTitle=False
):
assert isinstance(height, (type(None), int))
return core.Row(
panels=panels, height=height, title=title, showTitle=showTitle,
titleSize='h6', editable=False,
)
def REDRow(namespace, graphName, job, rule_root="job:", metric_root="request", extra_conditions="", collapse=False):
return G.Row(
title='%s QPS & Latency' % (graphName, ),
collapse=collapse,
panels=[
QPSGraph(namespace, graphName, job, metric_root, extra_conditions),
LatencyGraph(namespace, graphName, job, rule_root, metric_root, extra_conditions),
]
)
def _row(title):
return core.Row(panels=[
core.Graph(title=title,
dataSource='prometheus',
targets=[
core.Target(
expr=title,
legendFormat='{{namespace}}',
),
],
yAxes=[
core.YAxis(format=core.NO_FORMAT),
core.YAxis(format=core.SHORT_FORMAT),
])
])
def test_auto_refids():
"""
auto_ref_ids() provides refIds for all targets without refIds already
set.
"""
dashboard = G.Dashboard(
title="Test dashboard",
rows=[
G.Row(panels=[
G.Graph(
title="CPU Usage by Namespace (rate[5m])",
targets=[G.Target(expr="metric %d" % i)
for i in range(53)],
).auto_ref_ids()
]),
],
)
assert dashboard.rows[0].panels[0].targets[0].refId == 'A'
assert dashboard.rows[0].panels[0].targets[25].refId == 'Z'
assert dashboard.rows[0].panels[0].targets[26].refId == 'AA'
assert dashboard.rows[0].panels[0].targets[51].refId == 'AZ'
assert dashboard.rows[0].panels[0].targets[52].refId == 'BA'
def dashboard():
PROMETHEUS = "prometheus"
return G.Dashboard(
title="S4",
rows=[
G.Row(panels=[
G.Graph(
title="Signups",
dataSource=PROMETHEUS,
xAxis=X_TIME,
yAxes=[
G.YAxis(
format="none",
label="Count",
),
G.YAxis(
format="none",
label="Count",
),
],
targets=[
G.Target(
expr='wormhole_signup_started{pod=~"s4-signup.*"}',
legendFormat="Wormhole Signups Started",
refId="A",
),
G.Target(
expr='wormhole_signup_success{pod=~"s4-signup.*"}',
legendFormat="Wormhole Signups Completed",
refId="B",
),
G.Target(
expr='wormhole_signup_failure{pod=~"s4-signup.*"}',
legendFormat="Wormhole Signups Failed",
refId="C",
),
],
),
G.Graph(
title="Usage",
dataSource=PROMETHEUS,
# Stack the connection graphs on each other, revealing
# both a total and a distribution across different grid
# router instances.
stack=True,
tooltip=G.Tooltip(
valueType=G.INDIVIDUAL,
),
xAxis=X_TIME,
yAxes=[
G.YAxis(
format="none",
label="Count",
),
G.YAxis(
format="none",
label="Count",
),
],
targets=[
G.Target(
expr="grid_router_connections",
legendFormat="Tahoe-LAFS Connections",
refId="D",
),
],
),
]),
G.Row(
title="Cluster",
panels=[
cpu_usage(PROMETHEUS, ["1m", "5m", "10m"]),
memory_usage(PROMETHEUS),
network_usage(PROMETHEUS),
filesystem_usage(PROMETHEUS),
],
),
G.Row(panels=[
G.SingleStat(
title='Current Customer Deployments',
dataSource='prometheus',
valueName='current',
sparkline=G.SparkLine(show=True),
targets=[
G.Target(
expr='s4_deployment_gauge',
refId="E",
),
],
),
G.SingleStat(
title='Unhandled Errors',
dataSource='prometheus',
valueName='current',
sparkline=G.SparkLine(show=True),
targets=[
G.Target(
expr='s4_unhandled_error_counter',
refId="F",
),
],
),
]),
],
).auto_panel_ids()
yAxes=g.single_y_axis(format=g.SHORT_FORMAT, logBase=10),
),
]
# The final dashboard must be named 'dashboard' so that grafanalib will find it.
dashboard = g.Dashboard(
title="Master dashboard",
time=g.Time("now-30d", "now"),
templating=g.Templating(list=[
# Make it possible to use $source as a source.
g.Template(name="source", type="datasource", query="prometheus")
]),
rows=[
g.Row(
title="Clusterloader",
height=DEFAULT_PANEL_HEIGHT,
panels=CLUSTERLOADER_PANELS,
),
g.Row(
title="Overall cluster health",
height=DEFAULT_PANEL_HEIGHT,
panels=HEALTH_PANELS,
),
g.Row(title="etcd", height=DEFAULT_PANEL_HEIGHT, panels=ETCD_PANELS),
g.Row(title="kube-apiserver",
height=DEFAULT_PANEL_HEIGHT,
panels=APISERVER_PANELS),
g.Row(
title="kube-controller-manager",
height=DEFAULT_PANEL_HEIGHT,
panels=[
"3xx": BLUE,
"4xx": ORANGE,
"5xx": RED,
"success": GREEN,
"error": RED,
}
dashboard = G.Dashboard(
title='UCDAPI',
rows=[
G.Row(panels=[
G.SingleStat(
title='Pods Up',
dataSource='prometheus',
valueName='current',
sparkline=G.SparkLine(show=True),
targets=[
G.Target(
expr='count by(service) (up{service="ucdapi"} == 1)', )
]),
]),
G.Row(panels=[
G.Graph(
title='HTTP RPS',
dataSource='prometheus',
targets=[
G.Target(
expr=
'service_status:http_request_duration_seconds_count:irate{service="ucdapi",status_code=~"1.."}',
legendFormat='1xx',
refId='A'),
def run(self):
templateList = []
if self.options.use_sdp:
templateList.append(
G.Template(default="1",
dataSource="default",
name="sdpinst",
label="SDPInstance",
query="label_values(sdpinst)"))
if self.options.customer:
templateList.append(
G.Template(default="1",
dataSource="default",
name="customer",
label="Customer",
query="label_values(customer)"))
templateList.append(
G.Template(default="",
dataSource="default",
name="serverid",
label="ServerID",
query="label_values(serverid)"))
dashboard = G.Dashboard(title=self.options.title,
templating=G.Templating(list=templateList))
for metric in metrics:
if 'section' in metric:
dashboard.rows.append(
G.Row(title=metric['section'], showTitle=True))
continue
if 'row' in metric:
dashboard.rows.append(G.Row(title='', showTitle=False))
continue
if 'type' in metric and metric['type'] == 'gauge':
pass
# text = G.Text(title=metric['title'],
# dataSource='default')
# dashboard.rows[-1].panels.append(G.Text)
else:
yAxis = G.single_y_axis(format="short")
if 'yformat' in metric:
yAxis = G.single_y_axis(format=metric['yformat'])
graph = G.Graph(title=metric['title'],
dataSource='default',
maxDataPoints=1000,
legend=G.Legend(show=True,
alignAsTable=True,
min=True,
max=True,
avg=True,
current=True,
total=True,
sort='max',
sortDesc=True),
yAxes=yAxis)
refId = 'A'
for targ in metric['target']:
texp = targ['expr']
legend = "instance {{instance}}, serverid {{serverid}}"
if 'legend' in targ:
legend += ' %s' % targ['legend']
# Remove SDP
if not self.options.use_sdp:
texp = texp.replace('sdpinst="$sdpinst",', '')
if self.options.customer:
texp = texp.replace('{', '{customer="$customer",')
graph.targets.append(
G.Target(expr=texp, legendFormat=legend, refId=refId))
refId = chr(ord(refId) + 1)
dashboard.rows[-1].panels.append(graph)
# Auto-number panels - returns new dashboard
dashboard = dashboard.auto_panel_ids()
s = io.StringIO()
write_dashboard(dashboard, s)
print("""{
"dashboard": %s
}
""" % s.getvalue())
G.Row(
title="Retrieval Stats",
collapse=True,
panels=[
common.PromGraph(
title="Retrieval sent batches",
expressions=[
('{{url}}',
'sum(rate(prometheus_remote_storage_sent_batch_duration_seconds_count[1m])) by (url)'
),
],
),
common.PromGraph(
title="Retrieval batch latency",
expressions=[
('{{url}} 99th',
'histogram_quantile(0.99, sum(rate(prometheus_remote_storage_sent_batch_duration_seconds_bucket[2m])) by (url, le)) * 1e3'
),
('{{url}} 50th',
'histogram_quantile(0.50, sum(rate(prometheus_remote_storage_sent_batch_duration_seconds_bucket[2m])) by (url, le)) * 1e3'
),
('{{url}} mean',
'(sum(rate(prometheus_remote_storage_sent_batch_duration_seconds_sum[2m])) by (url) / sum(rate(prometheus_remote_storage_sent_batch_duration_seconds_count[2m])) by (url)) * 1e3'
),
],
yAxes=common.LATENCY_AXES,
),
common.PromGraph(
title="Retrieval sent samples",
expressions=[
('{{url}} success',
'sum(rate(prometheus_remote_storage_succeeded_samples_total[1m])) by (url)'
),
('{{url}} dropped',
'sum(rate(prometheus_remote_storage_dropped_samples_total[1m])) by (url)'
),
('{{url}} retried',
'sum(rate(prometheus_remote_storage_retried_samples_total[1m])) by (url)'
),
('{{url}} failure',
'sum(rate(prometheus_remote_storage_failed_samples_total[1m])) by (url)'
),
],
),
common.PromGraph(
title="Queue",
expressions=[
('{{url}}: queue length',
'sum(prometheus_remote_storage_pending_samples) by (url)'
),
('{{url}}: lag',
'max(time()-prometheus_remote_storage_queue_highest_sent_timestamp_seconds) by (url)'
),
('{{url}}: shards',
'max(prometheus_remote_storage_shards) by (url)'),
],
),
],
),
G.Row(title="Service status",
panels=[
G.SingleStat(
editable=True,
colorBackground=True,
colorValue=False,
colors=[G.ORANGE, G.GREEN, G.RED],
dataSource="Prometheus",
gauge=G.Gauge(
maxValue=100,
minValue=0,
show=False,
thresholdLabels=False,
thresholdMarkers=True,
),
title=app["name"],
targets=[
G.Target(
expr='up{{instance="{hostname}:{port}"}}'.format(
hostname=app["hostname"], port=app["port"]),
intervalFactor=2,
legendFormat="",
step=4,
)
],
span=2,
thresholds="0.1",
transparent=True,
valueMaps=[
G.ValueMap(op="=", text="N/A", value="null"),
G.ValueMap(op="=", text="Down", value="0"),
G.ValueMap(op="=", text="Up", value="1"),
],
sparkline=G.SparkLine(
fillColor=G.RGBA(31, 118, 109, 0.18),
full=False,
lineColor=G.RGB(31, 120, 193),
show=False,
),
rangeMaps=[G.RangeMap(start=None, text="N/A", end=None)],
valueName="current",
) for app in APPS
])
import common
dashboard = common.Dashboard(
uid='chunks',
title="Cortex > Chunks",
rows=[
G.Row(
panels=[
common.PromGraph(
title="Number of chunks (in memory, in ingesters)",
expressions=[
('', 'sum(cortex_ingester_memory_chunks{job="cortex/ingester"})'),
],
),
common.PromGraph(
title="Chunks per series",
expressions=[
(
'',
'sum(cortex_ingester_memory_chunks{job="cortex/ingester"}) / sum(cortex_ingester_memory_series{job="cortex/ingester"})'
),
],
),
]
),
G.Row(
panels=[
common.PromGraph(
title="Chunk Size Bytes (on flush)",
expressions=[
(
G.Row(
title="Configs",
collapse=True,
panels=[
common.PromGraph(
title="Known Configurations",
expressions=[
("Configurations",
'max(cortex_configs{job="cortex/ruler"})'),
("{{status}}",
'max by(status)(cortex_alertmanager_configs{job="cortex/alertmanager"})'
),
],
),
common.QPSGraph('cortex_configs', 'Configs', 'cortex/ruler'),
common.PromGraph(
title="Configs Latency",
expressions=[
("99th centile",
'histogram_quantile(0.99, sum(rate(cortex_configs_request_duration_seconds_bucket{job="cortex/ruler"}[2m])) by (le)) * 1e3'
),
("50th centile",
'histogram_quantile(0.50, sum(rate(cortex_configs_request_duration_seconds_bucket{job="cortex/ruler"}[2m])) by (le)) * 1e3'
),
("Mean",
'sum(rate(cortex_configs_request_duration_seconds_sum{job="cortex/ruler"}[2m])) / sum(rate(cortex_configs_request_duration_seconds_count{job="cortex/ruler"}[2m])) * 1e3'
),
],
yAxes=common.LATENCY_AXES,
),
]),
uid='am',
title="Cortex > Alertmanager",
rows=[
G.Row(
title='Operations',
panels=[
common.PromGraph(
title="Alerts",
expressions=[
("{{instance}} {{status}}",
'sum by (instance, status)(rate(alertmanager_alerts_received_total{job="cortex/alertmanager"}[2m]))'
),
("{{instance}} invalid",
'sum by (instance, status)(rate(alertmanager_alerts_invalid_total{job="cortex/alertmanager"}[2m]))'
),
],
yAxes=common.OPS_AXIS,
),
common.PromGraph(
title="Notifications",
expressions=[
("{{integration}}",
'sum by (integration)(rate(alertmanager_notifications_total{job="cortex/alertmanager"}[2m]))'
),
],
yAxes=common.OPS_AXIS,
),
]),
G.Row(
title='Alertmanager fetching configs',
collapse=True,
def dashboard():
PROMETHEUS = "prometheus"
return G.Dashboard(
title="S4",
rows=[
G.Row(panels=[
G.Graph(
title="Signups",
dataSource=PROMETHEUS,
xAxis=X_TIME,
yAxes=[
G.YAxis(
format="none",
label="Count",
),
G.YAxis(
format="none",
label="Count",
),
],
targets=[
G.Target(
# Filter down to just the signup pod since that's
# the only one where this metric value is
# meaningful. Some other pods report a 0 value
# for this metric because they happen to import
# the Python code that defines the object
# representing it.
#
# Also, sum over the selected series to account
# for pod replacement.
expr='sum(wormhole_signup_started{pod=~"s4-signup.*"})',
legendFormat="Wormhole Signups Started",
refId="A",
),
G.Target(
expr='sum(wormhole_signup_success{pod=~"s4-signup.*"})',
legendFormat="Wormhole Signups Completed",
refId="B",
),
G.Target(
expr='sum(wormhole_signup_failure{pod=~"s4-signup.*"})',
legendFormat="Wormhole Signups Failed",
refId="C",
),
],
),
G.Graph(
title="Usage",
dataSource=PROMETHEUS,
# Stack the connection graphs on each other, revealing
# both a total and a distribution across different grid
# router instances.
stack=True,
tooltip=G.Tooltip(
valueType=G.INDIVIDUAL,
),
xAxis=X_TIME,
yAxes=[
G.YAxis(
format="none",
label="Count",
),
G.YAxis(
format="none",
label="Count",
),
],
targets=[
G.Target(
expr="grid_router_connections",
legendFormat="Tahoe-LAFS Connections",
refId="D",
),
],
),
last_convergence(PROMETHEUS),
]),
G.Row(
title="Cluster",
panels=[
cpu_usage(PROMETHEUS, ["1m", "5m", "10m"]),
memory_usage(PROMETHEUS),
network_usage(PROMETHEUS),
filesystem_usage(PROMETHEUS),
],
),
G.Row(
title="Cluster2",
panels=[
process_open_fds(PROMETHEUS),
],
),
G.Row(panels=[
tahoe_lafs_transfer_rate(PROMETHEUS),
s4_customer_deployments(PROMETHEUS),
unhandled_errors(PROMETHEUS),
]),
],
).auto_panel_ids()
common.REDRow('cortex', 'Querier read', 'cortex/querier'),
G.Row(
title="Ingester",
panels=[
common.PromGraph(
title="Ingester read QPS",
expressions=[
('{{route}}: {{status_code}}',
'sum(rate(cortex_request_duration_seconds_count{job="cortex/ingester", route!="/cortex.Ingester/Push"}[1m])) by (route, status_code)'
),
],
yAxes=common.OPS_AXIS,
),
common.PromGraph(
title="Ingester read latency",
expressions=[
('{{route}}: 99th centile',
'job_route:cortex_request_duration_seconds:99quantile{job="cortex/ingester", route!="/cortex.Ingester/Push"} * 1e3'
),
('{{route}}: 50th centile',
'job_route:cortex_request_duration_seconds:50quantile{job="cortex/ingester", route!="/cortex.Ingester/Push"} * 1e3'
),
('{{route}}: Mean',
'sum(rate(cortex_request_duration_seconds_sum{job="cortex/ingester", route!="/cortex.Ingester/Push"}[2m])) by (route) * 1e3 / sum(rate(cortex_request_duration_seconds_count{job="cortex/ingester", route!="/cortex.Ingester/Push"}[2m])) by (route)'
),
],
yAxes=common.LATENCY_AXES,
),
],
),
G.Row(
def make(prefix, title):
def target(expr, **kw):
return G.Target(expr=expr.format(prefix), **kw)
return G.Dashboard(
title=title,
rows=[
G.Row(panels=[
G.SingleStat(
title='Pods up (web)',
dataSource='prometheus',
valueName='current',
sparkline=G.SparkLine(show=True),
targets=[
target(
expr=
'count by(service) (up{{service="{}-isaacranks-web"}} == 1)'
)
]),
G.SingleStat(
title='Pods up (rebuild)',
dataSource='prometheus',
valueName='current',
sparkline=G.SparkLine(show=True),
targets=[
target(
expr=
'count by(service) (up{{service="{}-isaacranks-rebuild"}} == 1)'
)
]),
]),
G.Row(panels=[
G.Graph(
title='HTTP RPS',
dataSource='prometheus',
targets=[
target(
expr=
'service_status:http_request_duration_seconds_count:irate{{service="{}-isaacranks-web",status_code=~"1.."}}',
legendFormat='1xx',
refId='A'),
target(
expr=
'service_status:http_request_duration_seconds_count:irate{{service="{}-isaacranks-web",status_code=~"2.."}}',
legendFormat='2xx',
refId='B'),
target(
expr=
'service_status:http_request_duration_seconds_count:irate{{service="{}-isaacranks-web",status_code=~"3.."}}',
legendFormat='3xx',
refId='C'),
target(
expr=
'service_status:http_request_duration_seconds_count:irate{{service="{}-isaacranks-web",status_code=~"4.."}}',
legendFormat='4xx',
refId='D'),
target(
expr=
'service_status:http_request_duration_seconds_count:irate{{service="{}-isaacranks-web",status_code=~"5.."}}',
legendFormat='5xx',
refId='E'),
],
aliasColors=ALIAS_COLORS,
yAxes=[
G.YAxis(format=G.OPS_FORMAT),
G.YAxis(format=G.SHORT_FORMAT, show=False)
],
nullPointMode=G.NULL_AS_ZERO,
stack=True,
lineWidth=0,
fill=10,
tooltip=G.Tooltip(valueType=G.INDIVIDUAL)),
G.Graph(
title='HTTP latency',
dataSource='prometheus',
targets=[
target(
expr=
'service:http_request_duration_seconds:50p{{service="{}-isaacranks-web"}} * 1000',
legendFormat='0.5q',
refId='A'),
target(
expr=
'service:http_request_duration_seconds:90p{{service="{}-isaacranks-web"}} * 1000',
legendFormat='0.9q',
refId='B'),
target(
expr=
'service:http_request_duration_seconds:99p{{service="{}-isaacranks-web"}} * 1000',
legendFormat='0.99q',
refId='C'),
],
aliasColors=ALIAS_COLORS,
yAxes=[
G.YAxis(format=G.MILLISECONDS_FORMAT),
G.YAxis(format=G.SHORT_FORMAT, show=False)
]),
]),
G.Row(panels=[
G.Graph(
title='Ballots',
dataSource='prometheus',
targets=[
target(
expr=
'service_version:isaacranks_ballot_generation_seconds_count:irate{{service="{}-isaacranks-web"}}',
legendFormat='{{version}}',
refId='A')
],
yAxes=[
G.YAxis(format=G.OPS_FORMAT),
G.YAxis(format=G.SHORT_FORMAT, show=False),
],
nullPointMode=G.NULL_AS_ZERO,
stack=True,
lineWidth=0,
fill=10,
tooltip=G.Tooltip(valueType=G.INDIVIDUAL)),
G.Graph(
title='Ballot latency',
dataSource='prometheus',
targets=[
target(
expr=
'service:isaacranks_ballot_generation_seconds:50p{{service="{}-isaacranks-web"}} * 1000',
legendFormat='0.5q',
refId='A'),
target(
expr=
'service:isaacranks_ballot_generation_seconds:90p{{service="{}-isaacranks-web"}} * 1000',
legendFormat='0.9q',
refId='B'),
target(
expr=
'service:isaacranks_ballot_generation_seconds:99p{{service="{}-isaacranks-web"}} * 1000',
legendFormat='0.99q',
refId='C'),
],
yAxes=[
G.YAxis(format=G.MILLISECONDS_FORMAT),
G.YAxis(format=G.SHORT_FORMAT, show=False)
]),
]),
G.Row(panels=[
G.Graph(
title='Votes',
dataSource='prometheus',
targets=[
target(
expr=
'service_version:isaacranks_vote_casting_seconds_count:irate{{service="{}-isaacranks-web"}}',
legendFormat='{{version}}',
refId='A')
],
yAxes=[
G.YAxis(format=G.OPS_FORMAT),
G.YAxis(format=G.SHORT_FORMAT, show=False)
],
nullPointMode=G.NULL_AS_ZERO,
stack=True,
lineWidth=0,
fill=10,
tooltip=G.Tooltip(valueType=G.INDIVIDUAL)),
G.Graph(
title='Vote latency',
dataSource='prometheus',
targets=[
target(
expr=
'service:isaacranks_vote_casting_seconds:50p{{service="{}-isaacranks-web"}} * 1000',
legendFormat='0.5q',
refId='A'),
target(
expr=
'service:isaacranks_vote_casting_seconds:90p{{service="{}-isaacranks-web"}} * 1000',
legendFormat='0.9q',
refId='B'),
target(
expr=
'service:isaacranks_vote_casting_seconds:99p{{service="{}-isaacranks-web"}} * 1000',
legendFormat='0.99q',
refId='C'),
],
yAxes=[
G.YAxis(format=G.MILLISECONDS_FORMAT),
G.YAxis(format=G.SHORT_FORMAT, show=False)
]),
]),
G.Row(panels=[
G.Graph(
title='Time since last rebuild',
dataSource='prometheus',
targets=[
target(
expr=
'time() - (isaacranks_last_rebuild_timestamp{{service="{}-isaacranks-rebuild"}} != 0)',
legendFormat='Age')
],
legend=G.Legend(current=True),
yAxes=[
G.YAxis(format=G.SECONDS_FORMAT),
G.YAxis(format=G.SHORT_FORMAT, show=False)
]),
G.Graph(
title='Rebuild duration',
dataSource='prometheus',
targets=[
target(
expr=
'isaacranks_last_rebuild_duration_seconds{{service="{}-isaacranks-rebuild"}} != 0',
legendFormat='Duration')
],
legend=G.Legend(current=True),
yAxes=[
G.YAxis(format=G.SECONDS_FORMAT),
G.YAxis(format=G.SHORT_FORMAT, show=False)
]),
])
]).auto_panel_ids()
import common
dashboard = common.Dashboard(
uid='cortex-blocks',
title="Cortex > Blocks",
rows=[
G.Row(
title="Data",
panels=[
common.PromGraph(
title="Number of series in memory, in ingesters",
expressions=[
('',
'sum(cortex_ingester_memory_series{job="cortex/ingester"})'
),
],
),
common.PromGraph(
title="Head chunks",
expressions=[
('{{instance}}', 'cortex_ingester_tsdb_head_chunks'),
],
),
]),
G.Row(
title="Resources",
panels=[
common.PromGraph(
title="Memory Usage",
expressions=[
('{{pod}}',