def test_should_generate_deployment_graph(self):
name = "service-1"
cloudwatch_data_source = "prod"
cluster_name = "cluster-1"
grid_pos = GridPos(1, 2, 3, 4)
panel = generate_deployment_graph(
name=name,
cloudwatch_data_source=cloudwatch_data_source,
cluster_name=cluster_name,
grid_pos=grid_pos,
)
panel.should.be.a(TimeSeries)
panel.title.should.eql("Deployments")
panel.dataSource.should.eql(cloudwatch_data_source)
panel.targets.should.have.length_of(1)
panel.targets[0].should.eql(
CloudwatchMetricsTarget(
alias="Deployment",
namespace="ECS/ContainerInsights",
statistics=["Maximum"],
metricName="DeploymentCount",
dimensions={"ServiceName": name, "ClusterName": cluster_name},
)
)
panel.gridPos.should.eql(grid_pos)
def test_should_generate_res_count_graph_with_alert(self):
name = "service-1"
cloudwatch_data_source = "prod"
loadbalancer = "loadbalancer-1"
target_group = "target-group-1"
grid_pos = GridPos(1, 2, 3, 4)
notifications = ["foo", "bar", "baz"]
panel = generate_res_count_graph(
name=name,
cloudwatch_data_source=cloudwatch_data_source,
grid_pos=grid_pos,
loadbalancer=loadbalancer,
target_group=target_group,
notifications=notifications,
)
panel.alert.should.be.a(Alert)
panel.alert.message.should.eql("{} has 5XX errors".format(name))
panel.alert.alertConditions.should.have.length_of(1)
panel.alert.alertConditions.should.eql(
[
AlertCondition(
Target(refId="A"),
timeRange=TimeRange("15m", "now"),
evaluator=GreaterThan(0),
reducerType=RTYPE_MAX,
operator=OP_AND,
),
]
)
def test_should_generate_pending_count_with_alerts_graph(self):
name = "service-1"
cloudwatch_data_source = "prod"
cluster_name = "cluster-1"
grid_pos = GridPos(1, 2, 3, 4)
notifications = ["foo", "bar"]
panel = generate_pending_count_graph(
name=name,
cloudwatch_data_source=cloudwatch_data_source,
cluster_name=cluster_name,
grid_pos=grid_pos,
notifications=notifications,
)
panel.alert.should.be.a(Alert)
panel.alert.gracePeriod.should.eql("15m")
panel.alert.alertConditions.should.have.length_of(1)
panel.alert.alertConditions[0].should.eql(
AlertCondition(
Target(refId="A"),
timeRange=TimeRange("5m", "now"),
evaluator=GreaterThan(0),
reducerType=RTYPE_MAX,
operator=OP_AND,
)
)
def test_should_generate_mem_utilization_percentage_with_alerts_graph(self):
name = "service-1"
cloudwatch_data_source = "prod"
cluster_name = "cluster-1"
grid_pos = GridPos(1, 2, 3, 4)
notifications = ["foo", "bar", "baz"]
expected_alert_condition = AlertCondition(
Target(refId="A"),
timeRange=TimeRange("15m", "now"),
evaluator=GreaterThan(85),
reducerType=RTYPE_MAX,
operator=OP_AND,
)
panel = generate_mem_utilization_percentage_graph(
name=name,
cloudwatch_data_source=cloudwatch_data_source,
cluster_name=cluster_name,
grid_pos=grid_pos,
notifications=notifications,
)
panel.alert.should.be.a(Alert)
panel.alert.alertConditions.should.have.length_of(1)
panel.alert.alertConditions[0].should.eql(expected_alert_condition)
panel.alert.notifications.should.eql(notifications)
def test_should_generate_pending_count_graph(self):
name = "service-1"
cloudwatch_data_source = "prod"
cluster_name = "cluster-1"
grid_pos = GridPos(1, 2, 3, 4)
notifications = []
panel = generate_pending_count_graph(
name=name,
cloudwatch_data_source=cloudwatch_data_source,
cluster_name=cluster_name,
grid_pos=grid_pos,
notifications=notifications,
)
panel.should.be.a(Graph)
panel.title.should.eql("Pending Tasks")
panel.dataSource.should.eql(cloudwatch_data_source)
panel.targets.should.have.length_of(1)
panel.targets[0].should.eql(
CloudwatchMetricsTarget(
alias="Containers",
namespace="ECS/ContainerInsights",
statistics=["Maximum"],
metricName="PendingTaskCount",
dimensions={"ServiceName": name, "ClusterName": cluster_name},
refId="A",
)
)
panel.gridPos.should.eql(grid_pos)
def test_should_generate_req_count_graph(self):
cloudwatch_data_source = "prod"
loadbalancer = "loadbalancer-1"
target_group = "target-group-1"
grid_pos = GridPos(1, 2, 3, 4)
panel = generate_req_count_graph(
cloudwatch_data_source=cloudwatch_data_source,
grid_pos=grid_pos,
loadbalancer=loadbalancer,
target_group=target_group,
)
panel.should.be.a(Graph)
panel.title.should.eql("Requests")
panel.dataSource.should.eql(cloudwatch_data_source)
panel.targets.should.have.length_of(2)
def test_should_generate_error_logs_panel(self):
name = "service-1"
grid_pos = GridPos(1, 2, 3, 4)
elasticsearch_data_source = "es"
es_query = 'tag: "booking-api" AND log.level: [50 TO *]'
panel = generate_error_logs_panel(
lucene_query=es_query,
grid_pos=grid_pos,
elasticsearch_data_source=elasticsearch_data_source,
)
panel.should.be.a(Logs)
panel.title.should.eql("Error Logs")
panel.gridPos.should.eql(grid_pos)
panel.dataSource.should.eql(elasticsearch_data_source)
panel.targets.should.have.length_of(1)
panel.targets[0].query.should.eql(es_query)
def test_should_generate_mem_utilization_graph(self):
name = "service-1"
cloudwatch_data_source = "prod"
cluster_name = "cluster-1"
grid_pos = GridPos(1, 2, 3, 4)
panel = generate_mem_utilization_graph(
name=name,
cloudwatch_data_source=cloudwatch_data_source,
cluster_name=cluster_name,
grid_pos=grid_pos,
)
panel.should.be.a(Graph)
panel.title.should.eql("Memory Utilization")
panel.dataSource.should.eql(cloudwatch_data_source)
panel.targets.should.have.length_of(4)
panel.gridPos.should.eql(grid_pos)
def test_should_generate_running_count_stats_panel(self):
name = "service-1"
cloudwatch_data_source = "prod"
cluster_name = "cluster-1"
grid_pos = GridPos(1, 2, 3, 4)
expected_targets = [
CloudwatchMetricsTarget(
alias="Desired",
namespace="ECS/ContainerInsights",
statistics=["Maximum"],
metricName="DesiredTaskCount",
dimensions={"ServiceName": name, "ClusterName": cluster_name},
),
CloudwatchMetricsTarget(
alias="Pending",
namespace="ECS/ContainerInsights",
statistics=["Maximum"],
metricName="PendingTaskCount",
dimensions={"ServiceName": name, "ClusterName": cluster_name},
),
CloudwatchMetricsTarget(
alias="Running",
namespace="ECS/ContainerInsights",
statistics=["Maximum"],
metricName="RunningTaskCount",
dimensions={"ServiceName": name, "ClusterName": cluster_name},
refId="A",
),
]
panel = generate_running_count_stats_panel(
name=name,
cloudwatch_data_source=cloudwatch_data_source,
cluster_name=cluster_name,
grid_pos=grid_pos,
)
panel.should.be.a(Stat)
panel.title.should.eql("Task Count")
panel.dataSource.should.eql(cloudwatch_data_source)
panel.gridPos.should.eql(grid_pos)
panel.colorMode.should.eql("background")
panel.alignment.should.eql("center")
panel.targets.should.have.length_of(3)
panel.targets.should.eql(expected_targets)
def test_should_generate_desired_count_without_alerts_graph_when_desired_is_1(self):
name = "service-1"
cloudwatch_data_source = "prod"
cluster_name = "cluster-1"
grid_pos = GridPos(1, 2, 3, 4)
max = 1
notifications = ["foo", "bar"]
panel = generate_desired_count_graph(
name=name,
cloudwatch_data_source=cloudwatch_data_source,
cluster_name=cluster_name,
grid_pos=grid_pos,
max=max,
notifications=notifications,
)
panel.alert.should.eql(None)
def test_should_generate_cpu_utilization_graph(self):
name = "service-1"
cloudwatch_data_source = "prod"
cluster_name = "cluster-1"
grid_pos = GridPos(1, 2, 3, 4)
panel = generate_cpu_utilization_graph(
name=name,
cloudwatch_data_source=cloudwatch_data_source,
cluster_name=cluster_name,
grid_pos=grid_pos,
)
expected_targets = [
CloudwatchMetricsTarget(
alias="Min",
namespace="AWS/ECS",
statistics=["Minimum"],
metricName="CPUUtilization",
dimensions={"ServiceName": name, "ClusterName": cluster_name},
),
CloudwatchMetricsTarget(
alias="Avg",
namespace="AWS/ECS",
statistics=["Average"],
metricName="CPUUtilization",
dimensions={"ServiceName": name, "ClusterName": cluster_name},
),
CloudwatchMetricsTarget(
alias="Max",
namespace="AWS/ECS",
statistics=["Maximum"],
metricName="CPUUtilization",
dimensions={"ServiceName": name, "ClusterName": cluster_name},
),
]
panel.should.be.a(Graph)
panel.title.should.eql("CPU Utilization Percentage")
panel.dataSource.should.eql(cloudwatch_data_source)
panel.targets.should.have.length_of(3)
panel.targets.should.eql(expected_targets)
panel.gridPos.should.eql(grid_pos)
def generate_ecs_alb_service_dashboard(
name: str,
cluster_name: str,
cloudwatch_data_source: str,
notifications: List[str],
environment: str,
loadbalancer: str,
target_group: str,
elasticsearch_data_source: str,
lucene_query: str,
max: int,
*args,
**kwargs,
):
"""Generate ECS Service dashboard"""
tags = ["ecs", "ecs-service", "containers", "service", environment]
panels = [
RowPanel(
title="Summary",
gridPos=GridPos(1, 24, 0, 0),
),
generate_running_count_stats_panel(
name=name,
cluster_name=cluster_name,
cloudwatch_data_source=cloudwatch_data_source,
grid_pos=GridPos(8, 12, 0, 1),
),
generate_deployment_graph(
name=name,
cluster_name=cluster_name,
cloudwatch_data_source=cloudwatch_data_source,
grid_pos=GridPos(8, 12, 12, 1),
),
RowPanel(title="Capacity", gridPos=GridPos(1, 24, 0, 9)),
generate_running_count_graph(
name=name,
cluster_name=cluster_name,
max=max,
cloudwatch_data_source=cloudwatch_data_source,
grid_pos=GridPos(8, 8, 0, 10),
notifications=notifications,
),
generate_desired_count_graph(
name=name,
cluster_name=cluster_name,
max=max,
cloudwatch_data_source=cloudwatch_data_source,
grid_pos=GridPos(8, 8, 8, 10),
notifications=notifications,
),
generate_pending_count_graph(
name=name,
cluster_name=cluster_name,
cloudwatch_data_source=cloudwatch_data_source,
grid_pos=GridPos(8, 8, 16, 10),
notifications=notifications,
),
RowPanel(title="Utilization", gridPos=GridPos(1, 24, 0, 18)),
generate_cpu_utilization_graph(
name=name,
cluster_name=cluster_name,
cloudwatch_data_source=cloudwatch_data_source,
grid_pos=GridPos(8, 8, 0, 19),
),
generate_mem_utilization_graph(
name=name,
cluster_name=cluster_name,
cloudwatch_data_source=cloudwatch_data_source,
grid_pos=GridPos(8, 8, 8, 19),
),
generate_mem_utilization_percentage_graph(
name=name,
cluster_name=cluster_name,
cloudwatch_data_source=cloudwatch_data_source,
grid_pos=GridPos(8, 8, 16, 19),
notifications=notifications,
),
RowPanel(title="Requests and Responses", gridPos=GridPos(1, 24, 0,
27)),
generate_req_count_graph(
loadbalancer=loadbalancer,
target_group=target_group,
cloudwatch_data_source=cloudwatch_data_source,
grid_pos=GridPos(8, 12, 0, 28),
),
generate_res_count_graph(
name=name,
loadbalancer=loadbalancer,
target_group=target_group,
cloudwatch_data_source=cloudwatch_data_source,
grid_pos=GridPos(8, 12, 12, 28),
notifications=notifications,
),
]
if elasticsearch_data_source and lucene_query:
panels += [
RowPanel(title="Logs", gridPos=GridPos(1, 24, 0, 36)),
generate_helpful_resources_panel(lucene_query=lucene_query,
grid_pos=GridPos(8, 24, 0, 37)),
generate_error_logs_panel(
grid_pos=GridPos(24, 24, 0, 45),
elasticsearch_data_source=elasticsearch_data_source,
lucene_query=lucene_query,
),
]
return Dashboard(
title="{} {}".format("ECS Service:", name),
editable=EDITABLE,
tags=tags,
timezone=TIMEZONE,
sharedCrosshair=SHARED_CROSSHAIR,
panels=panels,
refresh=DEFAULT_REFRESH,
).auto_panel_ids()
template_list = [
Template(default="",
name="instance",
label="Machine",
query="label_values(instance)"),
chip_template,
]
dashboard = Dashboard(
title="Temperature",
templating=Templating(template_list),
panels=[
RowPanel(
title="New Row",
gridPos=GridPos(h=1, w=24, x=0, y=8),
),
Graph(
title="$chip",
dataSource="Prometheus",
targets=[
Target(expr=("sensors_temp_input{" +
'instance="$instance",chip="$chip"}'),
legendFormat="{{feature}}",
refId="A"),
],
repeat=Repeat("v", "chip"),
yAxes=YAxes(
YAxis(format=CELSIUS_FORMAT),
YAxis(format=SHORT_FORMAT),
),
def test_should_generate_res_count_graph(self):
name = "service-1"
cloudwatch_data_source = "prod"
loadbalancer = "loadbalancer-1"
target_group = "target-group-1"
grid_pos = GridPos(1, 2, 3, 4)
notifications = []
panel = generate_res_count_graph(
name=name,
cloudwatch_data_source=cloudwatch_data_source,
grid_pos=grid_pos,
loadbalancer=loadbalancer,
target_group=target_group,
notifications=notifications,
)
panel.should.be.a(Graph)
panel.title.should.eql("Responses")
panel.gridPos.should.eql(grid_pos)
panel.dataSource.should.eql(cloudwatch_data_source)
panel.targets.should.have.length_of(4)
panel.targets.should.eql(
[
CloudwatchMetricsTarget(
alias="2xx",
namespace="AWS/ApplicationELB",
statistics=["Sum"],
metricName="HTTPCode_Target_2XX_Count",
dimensions={
"LoadBalancer": loadbalancer,
"TargetGroup": target_group,
},
),
CloudwatchMetricsTarget(
alias="3xx",
namespace="AWS/ApplicationELB",
statistics=["Sum"],
metricName="HTTPCode_Target_3XX_Count",
dimensions={
"LoadBalancer": loadbalancer,
"TargetGroup": target_group,
},
),
CloudwatchMetricsTarget(
alias="4xx",
namespace="AWS/ApplicationELB",
statistics=["Sum"],
metricName="HTTPCode_Target_4XX_Count",
dimensions={
"LoadBalancer": loadbalancer,
"TargetGroup": target_group,
},
),
CloudwatchMetricsTarget(
alias="5xx",
namespace="AWS/ApplicationELB",
statistics=["Sum"],
metricName="HTTPCode_Target_5XX_Count",
dimensions={
"LoadBalancer": loadbalancer,
"TargetGroup": target_group,
},
refId="A",
),
]
)
description="Example dashboard using the Random Walk and default Prometheus datasource",
tags=[
'example'
],
timezone="browser",
panels=[
TimeSeries(
title="Random Walk",
dataSource='default',
targets=[
Target(
datasource='grafana',
expr='example',
),
],
gridPos=GridPos(h=8, w=16, x=0, y=0),
),
GaugePanel(
title="Random Walk",
dataSource='default',
targets=[
Target(
datasource='grafana',
expr='example',
),
],
gridPos=GridPos(h=4, w=4, x=17, y=0),
),
TimeSeries(
title="Prometheus http requests",
dataSource='prometheus',
from grafanalib.core import (
Dashboard,
Graph,
GridPos,
OPS_FORMAT,
RowPanel,
SHORT_FORMAT,
SqlTarget,
YAxes,
YAxis,
)
dashboard = Dashboard(
title="Random stats from SQL DB",
panels=[
RowPanel(title="New row", gridPos=GridPos(h=1, w=24, x=0, y=8)),
Graph(
title="Some SQL Queries",
dataSource="Your SQL Source",
targets=[
SqlTarget(
rawSql=
'SELECT date as "time", metric FROM example WHERE $__timeFilter("time")',
refId="A",
),
],
yAxes=YAxes(
YAxis(format=OPS_FORMAT),
YAxis(format=SHORT_FORMAT),
),
gridPos=GridPos(h=8, w=24, x=0, y=9),