def test_prometheus_exporter_basic_auth():
# The dpgen service just checks that basic auth is present, not correct
with run_service("dpgen",
environment={
"NUM_METRICS": 3,
"REQUIRE_BASIC_AUTH": "yes"
}) as dpgen_cont:
with Agent.run(
dedent(f"""
monitors:
- type: prometheus-exporter
host: {container_ip(dpgen_cont)}
port: 3000
intervalSeconds: 2
extraDimensions:
source: prometheus
""")) as agent:
assert ensure_always(
lambda: not has_datapoint(agent.fake_services,
dimensions={"source": "prometheus"}),
timeout_seconds=5
), "got prometheus datapoint without basic auth (test setup is wrong)"
agent.config["monitors"][0]["username"] = "******"
agent.config["monitors"][0]["password"] = "******"
agent.write_config()
assert wait_for(
p(has_datapoint,
agent.fake_services,
dimensions={"source": "prometheus"
})), "didn't get prometheus datapoint"
def test_postgresql_database_filter():
with run_service("postgres",
buildargs={"POSTGRES_VERSION": "11-alpine"},
environment=ENV,
print_logs=False) as postgres_cont:
host = container_ip(postgres_cont)
assert wait_for(p(tcp_socket_open, host, 5432),
60), "service didn't start"
with Agent.run(
dedent(f"""
monitors:
- type: postgresql
host: {host}
port: 5432
connectionString: "user=test_user password=test_pwd dbname=postgres sslmode=disable"
databases: ['*', '!postgres']
""")) as agent:
for metric in METADATA.default_metrics:
assert wait_for(
p(has_datapoint,
agent.fake_services,
metric_name=metric,
dimensions={"database": "dvdrental"})
), f"Didn't get default postgresql metric {metric} for database dvdrental"
assert ensure_always(lambda: not has_datapoint(
agent.fake_services, dimensions={"database": "postgres"}
)), f"Should not get metric for postgres default database"
def test_kubernetes_cluster_namespace_scope(agent_image, minikube,
k8s_namespace):
monitors = [{
"type": "kubernetes-cluster",
"kubernetesAPI": {
"authType": "serviceAccount"
},
"namespace": "good"
}]
with run_k8s_with_agent(
agent_image,
minikube,
monitors,
namespace=k8s_namespace,
yamls=[local_file("good-pod.yaml"),
local_file("bad-pod.yaml")],
) as [backend, _]:
assert wait_for(
p(has_datapoint,
backend,
dimensions={"kubernetes_namespace":
"good"})), "timed out waiting for good pod metrics"
assert ensure_always(lambda: not has_datapoint(
backend, dimensions={"kubernetes_namespace": "bad"})
), "got pod metrics from unspecified namespace"
def test_signalfx_metadata():
with Agent.run("""
procPath: /proc
etcPath: /etc
monitors:
- type: collectd/signalfx-metadata
persistencePath: /var/run/signalfx-agent
- type: collectd/cpu
- type: collectd/disk
- type: collectd/memory
""") as agent:
assert wait_for(
p(has_datapoint, agent.fake_services, "cpu.utilization",
{"plugin": "signalfx-metadata"}))
assert wait_for(
p(has_datapoint, agent.fake_services, "disk_ops.total",
{"plugin": "signalfx-metadata"}))
assert wait_for(
p(has_datapoint, agent.fake_services, "memory.utilization",
{"plugin": "signalfx-metadata"}))
assert ensure_always(
lambda:
not has_datapoint(agent.fake_services, "cpu.utilization_per_core",
{"plugin": "signalfx-metadata"}),
timeout_seconds=5,
)
assert not has_log_message(agent.output.lower(),
"error"), "error found in agent output!"
def has_all_pod_datapoints():
for name in pod_names:
if not has_datapoint(
agent.fake_services,
dimensions={"kubernetes_pod_name": name}):
return False
return True
def test_signalfx_metadata():
with run_agent("""
procPath: /proc
etcPath: /etc
monitors:
- type: collectd/signalfx-metadata
persistencePath: /var/run/signalfx-agent
- type: collectd/cpu
- type: collectd/disk
- type: collectd/memory
""") as [backend, get_output, _]:
assert wait_for(
p(has_datapoint, backend, "cpu.utilization",
{"plugin": "signalfx-metadata"}))
assert wait_for(
p(has_datapoint, backend, "disk_ops.total",
{"plugin": "signalfx-metadata"}))
assert wait_for(
p(has_datapoint, backend, "memory.utilization",
{"plugin": "signalfx-metadata"}))
assert ensure_always(
lambda: not has_datapoint(backend, "cpu.utilization_per_core",
{"plugin": "signalfx-metadata"}))
assert not has_log_message(get_output().lower(),
"error"), "error found in agent output!"
def test_endpoint_config_mapping():
with run_service("postgres",
environment=[
"POSTGRES_USER=test_user",
"POSTGRES_PASSWORD=test_pwd", "POSTGRES_DB=postgres"
]) as postgres_container:
with Agent.run(f"""
observers:
- type: docker
monitors:
- type: postgresql
host: {container_ip(postgres_container)}
connectionString: "user=test_user password=test_pwd dbname=postgres sslmode=disable"
port: 5432
dimensionTransformations:
database: db
""") as agent:
assert wait_for(
p(has_datapoint,
agent.fake_services,
dimensions={
"db": "dvdrental"
})), "Didn't get properly transformed dimension name"
assert not has_datapoint(agent.fake_services,
dimensions={"database": "dvdrental"})
def test_overlapping_filter_with_monitor_type():
"""
Test overlapping filters with different negation. Blacklist is favored
"""
with Agent.run(
"""
monitors:
- type: collectd/memory
- type: collectd/uptime
metricsToExclude:
- metricName: uptime
negated: true
monitorType: collectd/uptime
- metricName: uptime
monitorType: collectd/uptime
"""
) as agent:
assert wait_for(lambda: has_datapoint(agent.fake_services, "memory.used"))
assert wait_for(lambda: has_datapoint(agent.fake_services, "memory.free"))
assert ensure_always(lambda: not has_datapoint(agent.fake_services, "uptime"), 5)
def test_cpu_all():
agent = run_agent_verify_all_metrics(
"""
monitors:
- type: cpu
extraMetrics: ["*"]
""",
METADATA,
)
assert has_datapoint(agent.fake_services, metric_name="cpu.utilization_per_core", dimensions={"cpu": "0"})
def test_docker_container_stats():
with run_service("nginx",
cpu_period=100_000,
cpu_quota=10000,
cpu_shares=50,
mem_limit=20 * 1024 * 1024) as nginx_container:
with Agent.run("""
monitors:
- type: cgroups
extraMetrics: ['*']
""") as agent:
verify(agent, METADATA.all_metrics)
expected_cgroup = "/docker/" + nginx_container.id
assert has_datapoint(agent.fake_services,
metric_name="cgroup.cpu_shares",
value=50,
dimensions={"cgroup": expected_cgroup})
assert has_datapoint(
agent.fake_services,
metric_name="cgroup.cpu_cfs_period_us",
value=100_000,
dimensions={"cgroup": expected_cgroup},
)
assert has_datapoint(
agent.fake_services,
metric_name="cgroup.cpu_cfs_quota_us",
value=10000,
dimensions={"cgroup": expected_cgroup},
)
assert has_datapoint(
agent.fake_services,
metric_name="cgroup.memory_limit_in_bytes",
value=20 * 1024 * 1024,
dimensions={"cgroup": expected_cgroup},
)
def test_process_monitor_executable_filter():
proc = psutil.Process(os.getpid())
self_proc_exec = proc.exe()
with Agent.run(f"""
monitors:
- type: process
executables:
- {self_proc_exec}
""") as agent:
verify(agent, METADATA.all_metrics)
assert has_datapoint(agent.fake_services,
dimensions={"executable": self_proc_exec})
def test_process_monitor_process_name_filter():
proc = psutil.Process(os.getpid())
self_proc_name = proc.name()
with Agent.run(f"""
monitors:
- type: process
processes:
- {self_proc_name}
""") as agent:
verify(agent, METADATA.all_metrics)
assert has_datapoint(agent.fake_services,
dimensions={"command": self_proc_name})
def test_kubernetes_cluster_namespace_scope(k8s_cluster):
yamls = [SCRIPT_DIR / "good-pod.yaml", SCRIPT_DIR / "bad-pod.yaml"]
with k8s_cluster.create_resources(yamls):
config = """
monitors:
- type: kubernetes-cluster
kubernetesAPI:
authType: serviceAccount
namespace: good
"""
with k8s_cluster.run_agent(agent_yaml=config) as agent:
assert wait_for(
p(has_datapoint, agent.fake_services, dimensions={"kubernetes_namespace": "good"})
), "timed out waiting for good pod metrics"
assert ensure_always(
lambda: not has_datapoint(agent.fake_services, dimensions={"kubernetes_namespace": "bad"})
), "got pod metrics from unspecified namespace"
def test_does_not_set_hostname_on_monitor_if_not_host_specific():
with Agent.run(
"""
hostname: acmeinc.com
monitors:
- type: cpu
disableHostDimensions: true
- type: memory
"""
) as agent:
assert wait_for(
p(has_datapoint, agent.fake_services, dimensions={"host": "acmeinc.com"}, metric_name="memory.utilization")
), "Didn't get overridden hostname in datapoint"
assert wait_for(p(has_datapoint, agent.fake_services, metric_name="cpu.utilization")), "Didn't get cpu metric"
assert ensure_always(
lambda: not has_datapoint(
agent.fake_services, metric_name="cpu.utilization", dimensions={"host": "acmeinc.com"}
)
), "Got overridden hostname in cpu datapoint"
def test_kubernetes_cluster_namespace_scope(agent_image, minikube,
k8s_namespace):
yamls = [local_file("good-pod.yaml"), local_file("bad-pod.yaml")]
with minikube.create_resources(yamls, namespace=k8s_namespace):
config = """
monitors:
- type: kubernetes-cluster
kubernetesAPI:
authType: serviceAccount
namespace: good
"""
with minikube.run_agent(agent_image,
config=config,
namespace=k8s_namespace) as [_, backend]:
assert wait_for(
p(has_datapoint,
backend,
dimensions={"kubernetes_namespace": "good"
})), "timed out waiting for good pod metrics"
assert ensure_always(lambda: not has_datapoint(
backend, dimensions={"kubernetes_namespace": "bad"}
)), "got pod metrics from unspecified namespace"
def test_elasticsearch_included():
with run_elasticsearch(
environment={"cluster.name": "testCluster"}) as es_container:
host = container_ip(es_container)
config = f"""
monitors:
- type: collectd/elasticsearch
host: {host}
port: 9200
username: elastic
password: testing123
"""
with Agent.run(config) as agent:
verify(agent, METADATA.default_metrics - EXCLUDED)
assert has_datapoint(
agent.fake_services,
dimensions={
"plugin_instance": "testCluster",
"plugin": "elasticsearch",
"index": "twitter"
},
)
assert not has_log_message(agent.output.lower(),
"error"), "error found in agent output!"
def no_node_uid_dim():
return not has_datapoint(agent.fake_services, dimensions={"kubernetes_node_uid": node.metadata.uid})
def has_all_pod_datapoints():
for name in pod_names:
if not has_datapoint(
backend, dimensions={"kubernetes_pod_name": name}):
return False
return True
def test():
assert has_datapoint(agent.fake_services, metric_name=target_metric)
def test_pcf_nozzle():
firehose_envelopes = [
{
"batch": [
{
"timestamp": "1580228407476075606",
"source_id": "uaa",
"instance_id": "",
"deprecated_tags": {},
"tags": {
"deployment": "cf-389cbac3d7a2c6c990c8",
"index": "ba5499ed-129c-48f2-877c-e270e5bd2648",
"ip": "10.0.4.7",
"job": "control",
"origin": "uaa",
"product": "Small Footprint Pivotal Application Service",
"system_domain": "sys.industry.cf-app.com",
},
"gauge": {"metrics": {"vitals.jvm.cpu.load": {"unit": "gauge", "value": 0}}},
}
]
},
{
"batch": [
{
"timestamp": "1580228407476126130",
"source_id": "uaa",
"instance_id": "",
"deprecated_tags": {},
"tags": {
"deployment": "cf-389cbac3d7a2c6c990c8",
"index": "ba5499ed-129c-48f2-877c-e270e5bd2648",
"ip": "10.0.4.7",
"job": "control",
"origin": "uaa",
"product": "Small Footprint Pivotal Application Service",
"system_domain": "sys.industry.cf-app.com",
},
"gauge": {"metrics": {"vitals.jvm.thread.count": {"unit": "gauge", "value": 47}}},
}
]
},
{
"batch": [
{
"timestamp": "1580228407476264719",
"source_id": "uaa",
"instance_id": "",
"deprecated_tags": {},
"tags": {
"deployment": "cf-389cbac3d7a2c6c990c8",
"index": "ba5499ed-129c-48f2-877c-e270e5bd2648",
"ip": "10.0.4.7",
"job": "control",
"origin": "uaa",
"product": "Small Footprint Pivotal Application Service",
"system_domain": "sys.industry.cf-app.com",
},
"gauge": {"metrics": {"vitals.jvm.non-heap.init": {"unit": "gauge", "value": 7_667_712}}},
}
]
},
{
"batch": [
{
"timestamp": "1580428783743352757",
"source_id": "doppler",
"instance_id": "",
"deprecated_tags": {},
"tags": {
"deployment": "cf-389cbac3d7a2c6c990c8",
"direction": "egress",
"index": "ba5499ed-129c-48f2-877c-e270e5bd2648",
"ip": "10.0.4.7",
"job": "control",
"metric_version": "2.0",
"origin": "loggregator.doppler",
"product": "Small Footprint Pivotal Application Service",
"system_domain": "sys.industry.cf-app.com",
},
"counter": {"name": "dropped", "delta": "0", "total": "149000"},
},
{
"timestamp": "1580428783743352757",
"source_id": "712c7c06-62eb-4cd4-92e3-1a58683d1866",
"instance_id": "0",
"deprecated_tags": {},
"tags": {
"deployment": "cf-389cbac3d7a2c6c990c8",
"index": "ba5499ed-129c-48f2-877c-e270e5bd2648",
"ip": "10.0.4.7",
"job": "compute",
"origin": "rep",
"product": "Small Footprint Pivotal Application Service",
"system_domain": "sys.industry.cf-app.com",
},
"gauge": {"metrics": {"cpu": {"unit": "gauge", "value": 555}}},
},
]
},
{
"batch": [
{
"timestamp": "1580428783743496839",
"source_id": "doppler",
"instance_id": "",
"deprecated_tags": {},
"tags": {
"deployment": "cf-389cbac3d7a2c6c990c8",
"direction": "ingress",
"index": "ba5499ed-129c-48f2-877c-e270e5bd2648",
"ip": "10.0.4.7",
"job": "control",
"metric_version": "2.0",
"origin": "loggregator.doppler",
"product": "Small Footprint Pivotal Application Service",
"system_domain": "sys.industry.cf-app.com",
},
"counter": {"name": "dropped", "delta": "0", "total": "0"},
},
{
"timestamp": "1580428783744624100",
"source_id": "doppler",
"instance_id": "",
"deprecated_tags": {},
"tags": {
"deployment": "cf-389cbac3d7a2c6c990c8",
"index": "ba5499ed-129c-48f2-877c-e270e5bd2648",
"ip": "10.0.4.7",
"job": "control",
"metric_version": "2.0",
"origin": "loggregator.doppler",
"product": "Small Footprint Pivotal Application Service",
"system_domain": "sys.industry.cf-app.com",
},
"counter": {"name": "egress", "delta": "0", "total": "1075978016"},
},
{
"timestamp": "1580428783744924877",
"source_id": "doppler",
"instance_id": "",
"deprecated_tags": {},
"tags": {
"deployment": "cf-389cbac3d7a2c6c990c8",
"index": "ba5499ed-129c-48f2-877c-e270e5bd2648",
"ip": "10.0.4.7",
"job": "control",
"metric_version": "2.0",
"origin": "loggregator.doppler",
"product": "Small Footprint Pivotal Application Service",
"system_domain": "sys.industry.cf-app.com",
},
"counter": {"name": "egress", "delta": "7457", "total": "1075985473"},
},
{
"timestamp": "1580428783833603896",
"source_id": "system_metrics_agent",
"instance_id": "",
"deprecated_tags": {},
"tags": {
"deployment": "service-instance_a474d20d-9a64-4bac-993d-e0f644604083",
"index": "2cc60900-9c39-4ec3-80bb-2c1d22c10130",
"ip": "10.0.8.28",
"job": "mongodb-config-agent",
"origin": "system_metrics_agent",
"product": "Small Footprint Pivotal Application Service",
"system_domain": "sys.industry.cf-app.com",
},
"gauge": {"metrics": {"system_cpu_sys": {"unit": "Percent", "value": 0.315_324_026_576_374_3}}},
},
{
"timestamp": "1580428783833625031",
"source_id": "system_metrics_agent",
"instance_id": "",
"deprecated_tags": {},
"tags": {
"deployment": "service-instance_a474d20d-9a64-4bac-993d-e0f644604083",
"index": "2cc60900-9c39-4ec3-80bb-2c1d22c10130",
"ip": "10.0.8.28",
"job": "mongodb-config-agent",
"origin": "system_metrics_agent",
"product": "Small Footprint Pivotal Application Service",
"system_domain": "sys.industry.cf-app.com",
},
"gauge": {
"metrics": {
"system_disk_ephemeral_inode_percent": {"unit": "Percent", "value": 0.103_100_393_700_787_4}
}
},
},
]
},
]
with run_fake_rlp_gateway(firehose_envelopes) as [gateway_url, _], run_fake_uaa() as [uaa_url, _]:
with Agent.run(
f"""
disableHostDimensions: true
monitors:
- type: cloudfoundry-firehose-nozzle
uaaUrl: {uaa_url}
rlpGatewayUrl: {gateway_url}
uaaUser: myusername
uaaPassword: mypassword
extraMetrics:
- "*"
"""
) as agent:
expected_time_series = [
[
"uaa.vitals.jvm.non-heap.init",
{
"deployment": "cf-389cbac3d7a2c6c990c8",
"index": "ba5499ed-129c-48f2-877c-e270e5bd2648",
"ip": "10.0.4.7",
"job": "control",
"origin": "uaa",
"product": "Small Footprint Pivotal Application Service",
"source_id": "uaa",
"system_domain": "sys.industry.cf-app.com",
},
],
[
"uaa.vitals.jvm.cpu.load",
{
"source_id": "uaa",
"deployment": "cf-389cbac3d7a2c6c990c8",
"index": "ba5499ed-129c-48f2-877c-e270e5bd2648",
"ip": "10.0.4.7",
"job": "control",
"origin": "uaa",
"product": "Small Footprint Pivotal Application Service",
"system_domain": "sys.industry.cf-app.com",
},
],
[
"uaa.vitals.jvm.thread.count",
{
"source_id": "uaa",
"deployment": "cf-389cbac3d7a2c6c990c8",
"index": "ba5499ed-129c-48f2-877c-e270e5bd2648",
"ip": "10.0.4.7",
"job": "control",
"origin": "uaa",
"product": "Small Footprint Pivotal Application Service",
"system_domain": "sys.industry.cf-app.com",
},
],
[
"doppler.dropped",
{
"source_id": "doppler",
"deployment": "cf-389cbac3d7a2c6c990c8",
"direction": "egress",
"index": "ba5499ed-129c-48f2-877c-e270e5bd2648",
"ip": "10.0.4.7",
"job": "control",
"metric_version": "2.0",
"origin": "loggregator.doppler",
"product": "Small Footprint Pivotal Application Service",
"system_domain": "sys.industry.cf-app.com",
},
],
[
"doppler.dropped",
{
"source_id": "doppler",
"deployment": "cf-389cbac3d7a2c6c990c8",
"direction": "ingress",
"index": "ba5499ed-129c-48f2-877c-e270e5bd2648",
"ip": "10.0.4.7",
"job": "control",
"metric_version": "2.0",
"origin": "loggregator.doppler",
"product": "Small Footprint Pivotal Application Service",
"system_domain": "sys.industry.cf-app.com",
},
],
[
"rep.cpu",
{
"source_id": "712c7c06-62eb-4cd4-92e3-1a58683d1866",
"instance_id": "0",
"deployment": "cf-389cbac3d7a2c6c990c8",
"index": "ba5499ed-129c-48f2-877c-e270e5bd2648",
"ip": "10.0.4.7",
"job": "compute",
"origin": "rep",
"product": "Small Footprint Pivotal Application Service",
"system_domain": "sys.industry.cf-app.com",
},
],
[
"doppler.egress",
{
"source_id": "doppler",
"deployment": "cf-389cbac3d7a2c6c990c8",
"index": "ba5499ed-129c-48f2-877c-e270e5bd2648",
"ip": "10.0.4.7",
"job": "control",
"metric_version": "2.0",
"origin": "loggregator.doppler",
"product": "Small Footprint Pivotal Application Service",
"system_domain": "sys.industry.cf-app.com",
},
],
[
"system_metrics_agent.system_cpu_sys",
{
"source_id": "system_metrics_agent",
"deployment": "service-instance_a474d20d-9a64-4bac-993d-e0f644604083",
"index": "2cc60900-9c39-4ec3-80bb-2c1d22c10130",
"ip": "10.0.8.28",
"job": "mongodb-config-agent",
"origin": "system_metrics_agent",
"product": "Small Footprint Pivotal Application Service",
"system_domain": "sys.industry.cf-app.com",
},
],
[
"system_metrics_agent.system_disk_ephemeral_inode_percent",
{
"source_id": "system_metrics_agent",
"deployment": "service-instance_a474d20d-9a64-4bac-993d-e0f644604083",
"index": "2cc60900-9c39-4ec3-80bb-2c1d22c10130",
"ip": "10.0.8.28",
"job": "mongodb-config-agent",
"origin": "system_metrics_agent",
"product": "Small Footprint Pivotal Application Service",
"system_domain": "sys.industry.cf-app.com",
},
],
]
for metric_name, dimensions in expected_time_series:
assert wait_for(p(has_time_series, agent.fake_services, metric_name=metric_name, dimensions=dimensions))
assert has_datapoint(agent.fake_services, metric_type=sf_pbuf.GAUGE, metric_name="uaa.vitals.jvm.cpu.load")
assert has_datapoint(
agent.fake_services, metric_type=sf_pbuf.CUMULATIVE_COUNTER, metric_name="doppler.dropped"
)
