[pull] master from DataDog:master

.github/workflows/assign-issue.yml

@@ -17,8 +17,7 @@ jobs:
       contents: read
       issues: write
       actions: write
-      id-token: write
-    uses: DataDog/issue-triage-action/.github/workflows/issue-triage.yml@ee200860ba35ba6ed3ad565e918afb16811fe975 # v1.0.0
+    uses: DataDog/issue-triage-action/.github/workflows/issue-triage.yml@b39f0bc12abc52fe8aa70dc9b9353bf307a13219 # v1.0.1
     with:
       slack_map_path: .github/workflows/config/github_slack_map.yaml
       environment_name: issue-assigner

envoy/changelog.d/24237.added

@@ -0,0 +1 @@
+Add adaptive concurrency gradient controller metrics.

envoy/datadog_checks/envoy/metrics.py

@@ -387,6 +387,27 @@
     'envoy_http_rbac_denied': 'http.rbac_denied',
     'envoy_http_rbac_shadow_allowed': 'http.rbac_shadow_allowed',
     'envoy_http_rbac_shadow_denied': 'http.rbac_shadow_denied',
+    'envoy_http_adaptive_concurrency_gradient_controller_rq_blocked': (
+        'http.adaptive_concurrency.gradient_controller.rq_blocked'
+    ),
+    'envoy_http_adaptive_concurrency_gradient_controller_min_rtt_calculation_active': (
+        'http.adaptive_concurrency.gradient_controller.min_rtt_calculation_active'
+    ),
+    'envoy_http_adaptive_concurrency_gradient_controller_concurrency_limit': (
+        'http.adaptive_concurrency.gradient_controller.concurrency_limit'
+    ),
+    'envoy_http_adaptive_concurrency_gradient_controller_gradient': (
+        'http.adaptive_concurrency.gradient_controller.gradient'
+    ),
+    'envoy_http_adaptive_concurrency_gradient_controller_burst_queue_size': (
+        'http.adaptive_concurrency.gradient_controller.burst_queue_size'
+    ),
+    'envoy_http_adaptive_concurrency_gradient_controller_min_rtt_msecs': (
+        'http.adaptive_concurrency.gradient_controller.min_rtt_msecs'
+    ),
+    'envoy_http_adaptive_concurrency_gradient_controller_sample_rtt_msecs': (
+        'http.adaptive_concurrency.gradient_controller.sample_rtt_msecs'
+    ),
     'envoy_http_local_rate_limit_enabled': 'http.local_rate_limit_enabled',
     'envoy_http_local_rate_limit_enforced': 'http.local_rate_limit_enforced',
     'envoy_http_local_rate_limit_rate_limited': 'http.local_rate_limit_rate_limited',
@@ -4092,6 +4113,69 @@
         ),
         'method': 'monotonic_count',
     },
+    'http.adaptive_concurrency.gradient_controller.rq_blocked': {
+        'tags': (
+            ('stat_prefix',),
+            (),
+            (),
+            (),
+        ),
+        'method': 'monotonic_count',
+    },
+    'http.adaptive_concurrency.gradient_controller.min_rtt_calculation_active': {
+        'tags': (
+            ('stat_prefix',),
+            (),
+            (),
+            (),
+        ),
+        'method': 'gauge',
+    },
+    'http.adaptive_concurrency.gradient_controller.concurrency_limit': {
+        'tags': (
+            ('stat_prefix',),
+            (),
+            (),
+            (),
+        ),
+        'method': 'gauge',
+    },
+    'http.adaptive_concurrency.gradient_controller.gradient': {
+        'tags': (
+            ('stat_prefix',),
+            (),
+            (),
+            (),
+        ),
+        'method': 'gauge',
+    },
+    'http.adaptive_concurrency.gradient_controller.burst_queue_size': {
+        'tags': (
+            ('stat_prefix',),
+            (),
+            (),
+            (),
+        ),
+        'method': 'gauge',
+    },
+    'http.adaptive_concurrency.gradient_controller.min_rtt_msecs': {
+        'tags': (
+            ('stat_prefix',),
+            (),
+            (),
+            (),
+        ),
+        'method': 'gauge',
+    },
+    'http.adaptive_concurrency.gradient_controller.sample_rtt_msecs': {
+        'tags': (
+            ('stat_prefix',),
+            (),
+            (),
+            (),
+        ),
+        'method': 'gauge',
+    },
 }
 # fmt: on
 

envoy/metadata.csv

@@ -1029,3 +1029,11 @@ envoy.tls_inspector.sni.not_found.count,count,,,,[OpenMetrics V2] Total number o
 envoy.tls_inspector.bytes_processed.bucket,count,,,,[OpenMetrics V2] Records sizes which records the number of bytes the tls_inspector processed while analyzing for tls usage,0,envoy,,
 envoy.tls_inspector.bytes_processed.count,count,,,,[OpenMetrics V2] Count of records sizes which records the number of bytes the tls_inspector processed while analyzing for tls usage,0,envoy,,
 envoy.tls_inspector.bytes_processed.sum,count,,,,[OpenMetrics V2] Total sum of records sizes which records the number of bytes the tls_inspector processed while analyzing for tls usage,0,envoy,,
+envoy.http.adaptive_concurrency.gradient_controller.rq_blocked,count,,request,,[Legacy] Total requests blocked by the adaptive concurrency filter.,-1,envoy,,
+envoy.http.adaptive_concurrency.gradient_controller.rq_blocked.count,count,,request,,[OpenMetrics V2] Total requests blocked by the adaptive concurrency filter.,-1,envoy,,
+envoy.http.adaptive_concurrency.gradient_controller.min_rtt_calculation_active,gauge,,,,"[OpenMetrics V2] Set to 1 if the gradient controller is in the process of a minRTT calculation, and 0 otherwise.",0,envoy,,
+envoy.http.adaptive_concurrency.gradient_controller.concurrency_limit,gauge,,request,,[OpenMetrics V2] The current concurrency limit.,0,envoy,,
+envoy.http.adaptive_concurrency.gradient_controller.gradient,gauge,,,,"[OpenMetrics V2] The current gradient value multiplied by 1,000.",0,envoy,,
+envoy.http.adaptive_concurrency.gradient_controller.burst_queue_size,gauge,,request,,[OpenMetrics V2] The current headroom value in the concurrency limit calculation.,0,envoy,,
+envoy.http.adaptive_concurrency.gradient_controller.min_rtt_msecs,gauge,,millisecond,,[OpenMetrics V2] The current measured minimum round-trip time.,0,envoy,,
+envoy.http.adaptive_concurrency.gradient_controller.sample_rtt_msecs,gauge,,millisecond,,[OpenMetrics V2] The current measured sample round-trip time aggregate.,0,envoy,,

envoy/tests/common.py

@@ -398,6 +398,28 @@
 
 CONNECTION_LIMIT_STAT_PREFIX_TAG = 'stat_prefix:ingress_http'
 
+ADAPTIVE_CONCURRENCY_PROMETHEUS_COUNTER_METRIC = "http.adaptive_concurrency.gradient_controller.rq_blocked.count"
+
+ADAPTIVE_CONCURRENCY_PROMETHEUS_VALUES = {
+    ADAPTIVE_CONCURRENCY_PROMETHEUS_COUNTER_METRIC: 5,
+    "http.adaptive_concurrency.gradient_controller.min_rtt_calculation_active": 0,
+    "http.adaptive_concurrency.gradient_controller.concurrency_limit": 30,
+    "http.adaptive_concurrency.gradient_controller.gradient": 1000,
+    "http.adaptive_concurrency.gradient_controller.burst_queue_size": 22,
+    "http.adaptive_concurrency.gradient_controller.min_rtt_msecs": 100,
+    "http.adaptive_concurrency.gradient_controller.sample_rtt_msecs": 95,
+}
+
+ADAPTIVE_CONCURRENCY_PROMETHEUS_METRICS = list(ADAPTIVE_CONCURRENCY_PROMETHEUS_VALUES.keys())
+
+ADAPTIVE_CONCURRENCY_PROMETHEUS_STAT_PREFIX_TAG = 'stat_prefix:ingress_http'
+
+ADAPTIVE_CONCURRENCY_PROMETHEUS_GAUGE_VALUES = {
+    metric: value
+    for metric, value in ADAPTIVE_CONCURRENCY_PROMETHEUS_VALUES.items()
+    if metric != ADAPTIVE_CONCURRENCY_PROMETHEUS_COUNTER_METRIC
+}
+
 TLS_INSPECTOR_METRICS = [
     "tls_inspector.client_hello_too_large.count",
     "tls_inspector.tls.found.count",

envoy/tests/docker/api_v2/front-envoy.yaml

@@ -24,6 +24,22 @@ static_resources:
                 route:
                   cluster: service2
           http_filters:
+          - name: envoy.filters.http.adaptive_concurrency
+            typed_config:
+              "@type": type.googleapis.com/envoy.extensions.filters.http.adaptive_concurrency.v3.AdaptiveConcurrency
+              gradient_controller_config:
+                sample_aggregate_percentile:
+                  value: 90
+                concurrency_limit_params:
+                  concurrency_update_interval: 0.1s
+                min_rtt_calc_params:
+                  jitter:
+                    value: 10
+                  interval: 60s
+                  request_count: 50
+              enabled:
+                default_value: true
+                runtime_key: "adaptive_concurrency.enabled"
           - name: envoy.router
   clusters:
   - name: service1

envoy/tests/docker/api_v3/front-envoy.yaml

@@ -132,6 +132,22 @@ static_resources:
               grpc_service:
                 envoy_grpc:
                   cluster_name: ext-proc
+          - name: envoy.filters.http.adaptive_concurrency
+            typed_config:
+              "@type": type.googleapis.com/envoy.extensions.filters.http.adaptive_concurrency.v3.AdaptiveConcurrency
+              gradient_controller_config:
+                sample_aggregate_percentile:
+                  value: 90
+                concurrency_limit_params:
+                  concurrency_update_interval: 0.1s
+                min_rtt_calc_params:
+                  jitter:
+                    value: 10
+                  interval: 60s
+                  request_count: 50
+              enabled:
+                default_value: true
+                runtime_key: "adaptive_concurrency.enabled"
           - name: envoy.filters.http.router
             typed_config:
               "@type": type.googleapis.com/envoy.extensions.filters.http.router.v3.Router

envoy/tests/fixtures/legacy/adaptive_concurrency.txt

@@ -0,0 +1,7 @@
+http.ingress_http.adaptive_concurrency.gradient_controller.rq_blocked: 5
+http.ingress_http.adaptive_concurrency.gradient_controller.min_rtt_calculation_active: 1
+http.ingress_http.adaptive_concurrency.gradient_controller.concurrency_limit: 30
+http.ingress_http.adaptive_concurrency.gradient_controller.gradient: 1000
+http.ingress_http.adaptive_concurrency.gradient_controller.burst_queue_size: 22
+http.ingress_http.adaptive_concurrency.gradient_controller.min_rtt_msecs: 100
+http.ingress_http.adaptive_concurrency.gradient_controller.sample_rtt_msecs: 95

envoy/tests/fixtures/openmetrics/openmetrics.txt

@@ -1377,3 +1377,17 @@ envoy_tls_inspector_bytes_processed_bucket{le="3600000"} 0
 envoy_tls_inspector_bytes_processed_bucket{le="+Inf"} 0
 envoy_tls_inspector_bytes_processed_sum{} 0
 envoy_tls_inspector_bytes_processed_count{} 0
+# TYPE envoy_http_adaptive_concurrency_gradient_controller_rq_blocked counter
+envoy_http_adaptive_concurrency_gradient_controller_rq_blocked{envoy_http_conn_manager_prefix="ingress_http"} 5
+# TYPE envoy_http_adaptive_concurrency_gradient_controller_min_rtt_calculation_active gauge
+envoy_http_adaptive_concurrency_gradient_controller_min_rtt_calculation_active{envoy_http_conn_manager_prefix="ingress_http"} 0
+# TYPE envoy_http_adaptive_concurrency_gradient_controller_concurrency_limit gauge
+envoy_http_adaptive_concurrency_gradient_controller_concurrency_limit{envoy_http_conn_manager_prefix="ingress_http"} 30
+# TYPE envoy_http_adaptive_concurrency_gradient_controller_gradient gauge
+envoy_http_adaptive_concurrency_gradient_controller_gradient{envoy_http_conn_manager_prefix="ingress_http"} 1000
+# TYPE envoy_http_adaptive_concurrency_gradient_controller_burst_queue_size gauge
+envoy_http_adaptive_concurrency_gradient_controller_burst_queue_size{envoy_http_conn_manager_prefix="ingress_http"} 22
+# TYPE envoy_http_adaptive_concurrency_gradient_controller_min_rtt_msecs gauge
+envoy_http_adaptive_concurrency_gradient_controller_min_rtt_msecs{envoy_http_conn_manager_prefix="ingress_http"} 100
+# TYPE envoy_http_adaptive_concurrency_gradient_controller_sample_rtt_msecs gauge
+envoy_http_adaptive_concurrency_gradient_controller_sample_rtt_msecs{envoy_http_conn_manager_prefix="ingress_http"} 95

envoy/tests/legacy/common.py

@@ -90,3 +90,17 @@
 ]
 
 RBAC_METRICS = RBAC_ENFORCE_METRICS + RBAC_SHADOW_METRICS
+
+ADAPTIVE_CONCURRENCY_STAT_PREFIX_TAG = ['stat_prefix:ingress_http']
+
+ADAPTIVE_CONCURRENCY_METRIC_VALUES = {
+    "envoy.http.adaptive_concurrency.gradient_controller.rq_blocked": 5,
+    "envoy.http.adaptive_concurrency.gradient_controller.min_rtt_calculation_active": 1,
+    "envoy.http.adaptive_concurrency.gradient_controller.concurrency_limit": 30,
+    "envoy.http.adaptive_concurrency.gradient_controller.gradient": 1000,
+    "envoy.http.adaptive_concurrency.gradient_controller.burst_queue_size": 22,
+    "envoy.http.adaptive_concurrency.gradient_controller.min_rtt_msecs": 100,
+    "envoy.http.adaptive_concurrency.gradient_controller.sample_rtt_msecs": 95,
+}
+
+ADAPTIVE_CONCURRENCY_METRICS = list(ADAPTIVE_CONCURRENCY_METRIC_VALUES.keys())

envoy/tests/legacy/test_e2e.py

@@ -7,7 +7,7 @@
 from datadog_checks.dev.utils import get_metadata_metrics
 from datadog_checks.envoy import Envoy
 
-from .common import FLAVOR, HOST
+from .common import ADAPTIVE_CONCURRENCY_METRICS, ADAPTIVE_CONCURRENCY_STAT_PREFIX_TAG, FLAVOR, HOST
 
 pytestmark = [pytest.mark.e2e]
 
@@ -288,6 +288,11 @@ def test_e2e(dd_agent_check, exercise_envoy):
     for metric in METRICS:
         aggregator.assert_metric(metric)
 
+    for metric in ADAPTIVE_CONCURRENCY_METRICS:
+        aggregator.assert_metric(metric)
+        for tag in ADAPTIVE_CONCURRENCY_STAT_PREFIX_TAG:
+            aggregator.assert_metric_has_tag(metric, tag)
+
     if FLAVOR == 'api_v2':
         for metric in METRICS_V2:
             aggregator.assert_metric(metric)

envoy/tests/legacy/test_integration.py

@@ -7,7 +7,14 @@
 from datadog_checks.dev.utils import get_metadata_metrics
 from datadog_checks.envoy.metrics import METRIC_PREFIX, METRICS
 
-from .common import ENVOY_VERSION, EXT_AUTHZ_METRICS, INSTANCES, RBAC_METRICS
+from .common import (
+    ADAPTIVE_CONCURRENCY_METRICS,
+    ADAPTIVE_CONCURRENCY_STAT_PREFIX_TAG,
+    ENVOY_VERSION,
+    EXT_AUTHZ_METRICS,
+    INSTANCES,
+    RBAC_METRICS,
+)
 
 CHECK_NAME = 'envoy'
 UNIQUE_METRICS = EXT_AUTHZ_METRICS + RBAC_METRICS
@@ -53,6 +60,20 @@ def test_success(aggregator, check, dd_run_check):
     aggregator.assert_metrics_using_metadata(metadata_metrics)
 
 
+def test_adaptive_concurrency_metrics(aggregator, check, dd_run_check):
+    instance = INSTANCES['main']
+    c = check(instance)
+    dd_run_check(c)
+
+    # Guard against a vacuous pass: the controller needs real traffic to emit stats.
+    assert any(metric in aggregator.metric_names for metric in ADAPTIVE_CONCURRENCY_METRICS)
+
+    for metric in ADAPTIVE_CONCURRENCY_METRICS:
+        aggregator.assert_metric(metric)
+        for tag in ADAPTIVE_CONCURRENCY_STAT_PREFIX_TAG:
+            aggregator.assert_metric_has_tag(metric, tag)
+
+
 def test_metadata_integration(datadog_agent, check):
     instance = INSTANCES['main']
     c = check(instance)

envoy/tests/legacy/test_unit.py

@@ -12,6 +12,9 @@
 from datadog_checks.envoy.metrics import METRIC_PREFIX, METRICS
 
 from .common import (
+    ADAPTIVE_CONCURRENCY_METRIC_VALUES,
+    ADAPTIVE_CONCURRENCY_METRICS,
+    ADAPTIVE_CONCURRENCY_STAT_PREFIX_TAG,
     CONNECTION_LIMIT_METRICS,
     CONNECTION_LIMIT_STAT_PREFIX_TAG,
     ENVOY_VERSION,
@@ -351,3 +354,19 @@ def test_connection_limit_metrics(aggregator, fixture_path, mock_http_response,
             aggregator.assert_metric_has_tag(metric, tag, count=1)
 
     aggregator.assert_metrics_using_metadata(get_metadata_metrics())
+
+
+def test_adaptive_concurrency_metrics(aggregator, fixture_path, mock_http_response, check, dd_run_check):
+    instance = INSTANCES['main']
+    c = check(instance)
+
+    mock_http_response(file_path=fixture_path('./legacy/adaptive_concurrency.txt'))
+    dd_run_check(c)
+
+    # Pin the fixture values so a wrong mapping or metric type would be caught, not just a wrong name.
+    for metric in ADAPTIVE_CONCURRENCY_METRICS:
+        aggregator.assert_metric(metric, value=ADAPTIVE_CONCURRENCY_METRIC_VALUES[metric])
+        for tag in ADAPTIVE_CONCURRENCY_STAT_PREFIX_TAG:
+            aggregator.assert_metric_has_tag(metric, tag, count=1)
+
+    aggregator.assert_metrics_using_metadata(get_metadata_metrics())

envoy/tests/test_e2e.py

@@ -8,6 +8,7 @@
 from datadog_checks.envoy import Envoy
 
 from .common import (
+    ADAPTIVE_CONCURRENCY_PROMETHEUS_METRICS,
     CONNECTION_LIMIT_METRICS,
     DEFAULT_INSTANCE,
     FLAKY_METRICS,
@@ -24,7 +25,13 @@
 def test_e2e(dd_agent_check, exercise_envoy):
     aggregator = dd_agent_check(DEFAULT_INSTANCE, rate=True)
 
-    for metric in PROMETHEUS_METRICS + LOCAL_RATE_LIMIT_METRICS + CONNECTION_LIMIT_METRICS + TLS_INSPECTOR_METRICS:
+    for metric in (
+        PROMETHEUS_METRICS
+        + LOCAL_RATE_LIMIT_METRICS
+        + CONNECTION_LIMIT_METRICS
+        + TLS_INSPECTOR_METRICS
+        + ADAPTIVE_CONCURRENCY_PROMETHEUS_METRICS
+    ):
         formatted_metric = "envoy.{}".format(metric)
         if metric in FLAKY_METRICS:
             aggregator.assert_metric(formatted_metric, at_least=0)