cw-infra-apps/grafana/charts/feature-integrations/tests/tempo_metrics_test.yaml
2025-05-28 10:01:42 +03:00

359 lines
15 KiB
YAML

---
# yamllint disable rule:document-start rule:line-length rule:trailing-spaces rule:empty-lines
suite: Test Tempo Integration - Metrics
templates:
- configmap.yaml
tests:
- it: should create the tempo metrics config
set:
deployAsConfigMap: true
tempo:
instances:
- name: tempo
labelSelectors:
app.kubernetes.io/name: tempo
asserts:
- isKind:
of: ConfigMap
- equal:
path: data["metrics.alloy"]
value: |-
declare "tempo_integration" {
argument "metrics_destinations" {
comment = "Must be a list of metric destinations where collected metrics should be forwarded to"
}
declare "tempo_integration_discovery" {
argument "namespaces" {
comment = "The namespaces to look for targets in (default: [] is all namespaces)"
optional = true
}
argument "field_selectors" {
comment = "The field selectors to use to find matching targets (default: [])"
optional = true
}
argument "label_selectors" {
comment = "The label selectors to use to find matching targets (default: [\"app.kubernetes.io/name=tempo\"])"
optional = true
}
argument "port_name" {
comment = "The of the port to scrape metrics from (default: http-metrics)"
optional = true
}
// tempo service discovery for all of the pods
discovery.kubernetes "tempo_pods" {
role = "pod"
selectors {
role = "pod"
field = string.join(coalesce(argument.field_selectors.value, []), ",")
label = string.join(coalesce(argument.label_selectors.value, ["app.kubernetes.io/name=tempo"]), ",")
}
namespaces {
names = coalesce(argument.namespaces.value, [])
}
}
// tempo relabelings (pre-scrape)
discovery.relabel "tempo_pods" {
targets = discovery.kubernetes.tempo_pods.targets
// keep only the specified metrics port name, and pods that are Running and ready
rule {
source_labels = [
"__meta_kubernetes_pod_container_port_name",
"__meta_kubernetes_pod_phase",
"__meta_kubernetes_pod_ready",
"__meta_kubernetes_pod_container_init",
]
separator = "@"
regex = coalesce(argument.port_name.value, "http-metrics") + "@Running@true@false"
action = "keep"
}
// the tempo-mixin expects the job label to be namespace/component
rule {
source_labels = ["__meta_kubernetes_namespace","__meta_kubernetes_pod_label_app_kubernetes_io_component"]
separator = "/"
target_label = "job"
}
rule {
source_labels = ["__meta_kubernetes_namespace"]
target_label = "namespace"
}
rule {
source_labels = ["__meta_kubernetes_pod_name"]
target_label = "pod"
}
rule {
source_labels = ["__meta_kubernetes_pod_container_name"]
target_label = "container"
}
// set the workload to the controller kind and name
rule {
action = "lowercase"
source_labels = ["__meta_kubernetes_pod_controller_kind"]
target_label = "workload_type"
}
rule {
source_labels = ["__meta_kubernetes_pod_controller_name"]
target_label = "workload"
}
// remove the hash from the ReplicaSet
rule {
source_labels = [
"workload_type",
"workload",
]
separator = "/"
regex = "replicaset/(.+)-.+$"
target_label = "workload"
}
// set the app name if specified as metadata labels "app:" or "app.kubernetes.io/name:" or "k8s-app:"
rule {
action = "replace"
source_labels = [
"__meta_kubernetes_pod_label_app_kubernetes_io_name",
"__meta_kubernetes_pod_label_k8s_app",
"__meta_kubernetes_pod_label_app",
]
separator = ";"
regex = "^(?:;*)?([^;]+).*$"
replacement = "$1"
target_label = "app"
}
// set the component if specified as metadata labels "component:" or "app.kubernetes.io/component:" or "k8s-component:"
rule {
action = "replace"
source_labels = [
"__meta_kubernetes_pod_label_app_kubernetes_io_component",
"__meta_kubernetes_pod_label_k8s_component",
"__meta_kubernetes_pod_label_component",
]
regex = "^(?:;*)?([^;]+).*$"
replacement = "$1"
target_label = "component"
}
// set a source label
rule {
action = "replace"
replacement = "kubernetes"
target_label = "source"
}
}
export "output" {
value = discovery.relabel.tempo_pods.output
}
}
declare "tempo_integration_scrape" {
argument "targets" {
comment = "Must be a list() of targets"
}
argument "forward_to" {
comment = "Must be a list(MetricsReceiver) where collected metrics should be forwarded to"
}
argument "job_label" {
comment = "The job label to add for all Tempo metrics (default: integrations/tempo)"
optional = true
}
argument "keep_metrics" {
comment = "A regular expression of metrics to keep (default: see below)"
optional = true
}
argument "drop_metrics" {
comment = "A regular expression of metrics to drop (default: see below)"
optional = true
}
argument "scrape_interval" {
comment = "How often to scrape metrics from the targets (default: 60s)"
optional = true
}
argument "max_cache_size" {
comment = "The maximum number of elements to hold in the relabeling cache (default: 100000). This should be at least 2x-5x your largest scrape target or samples appended rate."
optional = true
}
argument "clustering" {
comment = "Whether or not clustering should be enabled (default: false)"
optional = true
}
prometheus.scrape "tempo" {
job_name = coalesce(argument.job_label.value, "integrations/tempo")
forward_to = [prometheus.relabel.tempo.receiver]
targets = argument.targets.value
scrape_interval = coalesce(argument.scrape_interval.value, "60s")
clustering {
enabled = coalesce(argument.clustering.value, false)
}
}
// tempo metric relabelings (post-scrape)
prometheus.relabel "tempo" {
forward_to = argument.forward_to.value
max_cache_size = coalesce(argument.max_cache_size.value, 100000)
// drop metrics that match the drop_metrics regex
rule {
source_labels = ["__name__"]
regex = coalesce(argument.drop_metrics.value, "")
action = "drop"
}
// keep only metrics that match the keep_metrics regex
rule {
source_labels = ["__name__"]
regex = coalesce(argument.keep_metrics.value, "(.+)")
action = "keep"
}
// the tempo-mixin expects the instance label to be the node name
rule {
source_labels = ["node"]
target_label = "instance"
replacement = "$1"
}
rule {
action = "labeldrop"
regex = "node"
}
// set the memcached exporter container name from container="exporter" to container="memcached"
rule {
source_labels = ["component", "container"]
separator = ";"
regex = "memcached-[^;]+;exporter"
target_label = "container"
replacement = "memcached"
}
}
}
tempo_integration_discovery "tempo" {
namespaces = []
label_selectors = ["app.kubernetes.io/name=tempo"]
port_name = "http-metrics"
}
tempo_integration_scrape "tempo" {
targets = tempo_integration_discovery.tempo.output
job_label = "integrations/tempo"
clustering = true
keep_metrics = "up|scrape_samples_scraped|container_cpu_usage_seconds_total|container_memory_working_set_bytes|container_network_receive_bytes_total|container_network_transmit_bytes_total|container_spec_cpu_period|container_spec_cpu_quota|container_spec_memory_limit_bytes|gauge_memberlist_health_score|go_gc_duration_seconds_count|go_goroutines|go_memstats_heap_inuse_bytes|kube_deployment_spec_replicas|kube_deployment_status_replicas_unavailable|kube_deployment_status_replicas_updated|kube_pod_container_info|kube_pod_container_resource_requests|kube_pod_container_status_restarts_total|kube_statefulset_replicas|kube_statefulset_status_replicas_current|kube_statefulset_status_replicas_ready|kube_statefulset_status_replicas_updated|kubelet_volume_stats_available_bytes|promtail_custom_bad_words_total|tempo_build_info|tempo_discarded_spans_total|tempo_distributor_bytes_received_total|tempo_distributor_metrics_generator_pushes_failures_total|tempo_distributor_push_duration_seconds_bucket|tempo_distributor_push_duration_seconds_count|tempo_distributor_push_duration_seconds_sum|tempo_distributor_queue_length|tempo_distributor_spans_received_total|tempo_ingester_blocks_cleared_total|tempo_ingester_blocks_flushed_total|tempo_ingester_failed_flushes_total|tempo_ingester_flush_duration_seconds_bucket|tempo_ingester_live_traces|tempo_ingester_traces_created_total|tempo_limits_defaults|tempo_limits_overrides|tempo_memberlist_client_cluster_members_count|tempo_memberlist_client_cluster_node_health_score|tempo_memberlist_client_kv_store_count|tempo_memcache_request_duration_seconds_bucket|tempo_memcache_request_duration_seconds_count|tempo_memcache_request_duration_seconds_sum|tempo_metrics_generator_bytes_received_total|tempo_metrics_generator_registry_active_series|tempo_metrics_generator_spans_discarded_total|tempo_metrics_generator_spans_received_total|tempo_query_frontend_queries_total|tempo_receiver_accepted_spans|tempo_receiver_refused_spans|tempo_request_duration_seconds_bucket|tempo_request_duration_seconds_count|tempo_request_duration_seconds_sum|tempo_vulture_trace_error_total|tempo_vulture_trace_total|tempodb_backend_request_duration_seconds_bucket|tempodb_backend_request_duration_seconds_count|tempodb_backend_request_duration_seconds_sum|tempodb_blocklist_length|tempodb_blocklist_poll_errors_total|tempodb_compaction_blocks_total|tempodb_compaction_bytes_written_total|tempodb_compaction_errors_total|tempodb_compaction_objects_combined_total|tempodb_compaction_objects_written_total|tempodb_compaction_outstanding_blocks|tempodb_retention_deleted_total|tempodb_retention_errors_total|tempodb_retention_marked_for_deletion_total|tempodb_work_queue_length|tempodb_work_queue_max"
scrape_interval = "60s"
max_cache_size = 100000
forward_to = argument.metrics_destinations.value
}
}
- it: should allow you to restrict the namespaces for metrics
set:
deployAsConfigMap: true
tempo:
instances:
- name: tempo
namespaces:
- k8smon
asserts:
- isKind:
of: ConfigMap
- matchRegex:
path: data["metrics.alloy"]
pattern: namespaces = \["k8smon"\]
- it: should allow you to disable the default allow list
set:
deployAsConfigMap: true
tempo:
instances:
- name: tempo
metrics:
tuning:
useDefaultAllowList: false
asserts:
- isKind:
of: ConfigMap
- matchRegex:
path: data["metrics.alloy"]
# The pattern should look like this, but since the regex is escaped, it will be a bit different
# tempo_integration_scrape "tempo" {
# targets = tempo_integration_discovery.tempo.output
# job_label = "integrations/tempo"
# clustering = true
# scrape_interval = "60s"
# max_cache_size = 100000
# forward_to = argument.metrics_destinations.value
# }
pattern: |-
\s*tempo_integration_scrape "tempo" {
\s* targets = tempo_integration_discovery.tempo.output
\s* job_label = "integrations/tempo"
\s* clustering = true
\s* scrape_interval = "60s"
\s* max_cache_size = 100000
\s* forward_to = argument.metrics_destinations.value
\s*}
- it: should allow you to specific which metrics to include
set:
deployAsConfigMap: true
tempo:
instances:
- name: tempo
metrics:
tuning:
useDefaultAllowList: false
includeMetrics:
- foo
- bar
asserts:
- isKind:
of: ConfigMap
- matchRegex:
path: data["metrics.alloy"]
pattern: keep_metrics = "up\|scrape_samples_scraped\|foo\|bar"
- it: should allow you to specific which metrics to exclude
set:
deployAsConfigMap: true
tempo:
instances:
- name: tempo
metrics:
tuning:
useDefaultAllowList: false
excludeMetrics:
- foo
- bar
asserts:
- isKind:
of: ConfigMap
- matchRegex:
path: data["metrics.alloy"]
pattern: drop_metrics = "foo\|bar"