Add root-level cluster_name and standardize collection_interval for OpenTelemetry features by Paamicky · Pull Request #2189 · aws/amazon-cloudwatch-agent

Paamicky · 2026-07-09T15:59:13Z

Description of the issue

The OpenTelemetry config section had two inconsistencies: cluster_name was required per-component, forcing customers to repeat the same cluster value for every feature, and the collection interval field name differed across features (metrics_collection_interval vs collection_interval). This has been updated to use the newer version collection_interval.

Description of changes

Root-level cluster_name: Moved to opentelemetry.cluster_name. A single value now applies to all features (Container Insights, Prometheus, Host Metrics, OTLP). Container Insights and Prometheus require it; Host Metrics and OTLP apply it when present.
Standardized collection_interval: All OpenTelemetry features now use collection_interval with a 30-second default. Host Metrics renamed from metrics_collection_interval. Database Insights retains its own hardcoded intervals.
Centralized validation: ClusterNameRegex moved to the common package. GetOtelClusterName and GetCollectionInterval added as shared helpers.
Schema updated: schema.json reflects the new root-level cluster_name field and standardized collection_interval across features.
Renamed mode to role in the Container Insights JSON config field (opentelemetry.collect.container_insights.role) to align with the CWAGENT_ROLE environment
variable naming.

Tests

Unit tests for GetOtelClusterName verifies root-level reads, env var is not used as fallback, empty config returns empty
Unit tests for GetCollectionInterval verifies feature-level override, 30s default, zero/negative values ignored
Updated golden file tests and confirm set_cluster_name processor appears in all applicable pipelines
All existing translator tests pass

License

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.

jefchien · 2026-07-10T18:16:42Z

+    transform/set_cluster_name:
+        error_mode: ignore
+        flatten_data: false
+        log_statements:
+            - context: resource
+              error_mode: ignore
+              statements:
+                - set(resource.attributes["k8s.cluster.name"], "TestCluster")
+        metric_statements: []
+        trace_statements: []


Shouldn't this have resource statements for each of the resource types? I see the same processor is being used in metric pipelines as well.

Fixed. otlp logs now uses its own processor (transform/set_cluster_name_logs) so it doesn't overwrite the metric_statements used by the metrics pipelines

Can't they share a single transform processor?

Fixed. set_cluster_name now registers with both metric_statements and log_statements in a single processor. Pipelines will use statements relevant to its signal.

WithTraceResourceStatements is not yet supported in the transform processor.

// Traces are not handled: WithTraceResourceStatements is not yet supported.
// Traces sent via OTLP on EKS should include k8s.cluster.name from the SDK.

This will need a separate PR if we want to add it.

Can you add WithTraceResourceStatements or a WithResourceStatements to that transformprocessor translator so it's consistent?

jefchien · 2026-07-13T21:38:23Z

+
+// GetCollectionInterval returns the collection interval for an OpenTelemetry feature.
+// It checks the feature-specific collection_interval field. Default is 30s.
+func GetCollectionInterval(conf *confmap.Conf, featureKey string) time.Duration {


This does the exact same thing as GetOrDefaultDuration

yeah removed to use the already existing function: GetOrDefaultDuration

jefchien · 2026-07-13T21:39:04Z

+
+// GetOtelClusterName returns the cluster name for OpenTelemetry-based features.
+// It reads exclusively from opentelemetry::cluster_name.
+func GetOtelClusterName(conf *confmap.Conf) string {


I don't think we need a separate function for this.

removed the function wrapper and defined inline at each call site.

jefchien · 2026-07-13T21:39:50Z

 }
+
+// DefaultCollectionInterval is the default collection interval for OpenTelemetry-based features.
+const DefaultCollectionInterval = 30 * time.Second


Don't define this here. It doesn't apply to all cases of collection interval. It's only for the opentelemetry pipelines.

Moved to a local constant in each package that uses it.

mitali-salvi · 2026-07-16T17:31:20Z

 		"opentelemetry": map[string]interface{}{
 			"collect": map[string]interface{}{
 				"container_insights": map[string]interface{}{
 					"cluster_name": "test-cluster",


seems like the cluster_name field is still showing up under container_insights section instead of showing up under opentelemetry/collect section, mind updating this please ?

…OpenTelemetry features

github-actions · 2026-07-17T14:57:45Z

Binary Size Report

linux/amd64

Binary	PR	vs `main` (`5d99671`)	vs `v1.300070.0`
amazon-cloudwatch-agent	233.3 MB	+0 B	${\color{red}▲}$ +2.0 MB (+0.8%)
amazon-cloudwatch-agent-config-wizard	2.0 MB	+0 B	${\color{red}▲}$ +86.0 KB (+4.5%)
config-downloader	2.0 MB	+0 B	${\color{red}▲}$ +86.0 KB (+4.5%)
config-translator	2.0 MB	+0 B	${\color{red}▲}$ +86.0 KB (+4.5%)
start-amazon-cloudwatch-agent	2.5 MB	+0 B	${\color{red}▲}$ +81.9 KB (+3.4%)
workload-discovery	3.0 MB	+0 B	${\color{red}▲}$ +86.0 KB (+2.9%)
Total	244.8 MB	+0 B	${\color{red}▲}$ +2.4 MB (+1.0%)

linux/amd64 amazon-cloudwatch-agent (last 9 main commits + this PR)

 233 ┤                                        
     ┤    ▄▄▄ ▄▄▄ ▄▄▄ ▄▄▄ ▄▄▄ ▄▄▄ ▄▄▄ ███ ███ 
     ┤    ███ ███ ███ ███ ███ ███ ███ ███ ███ 
     ┤    ███ ███ ███ ███ ███ ███ ███ ███ ███ 
     ┤    ███ ███ ███ ███ ███ ███ ███ ███ ███ 
     ┤    ███ ███ ███ ███ ███ ███ ███ ███ ███ 
     ┤    ███ ███ ███ ███ ███ ███ ███ ███ ███ 
     ┤    ███ ███ ███ ███ ███ ███ ███ ███ ███ 
 231 ┤▄▄▄ ███ ███ ███ ███ ███ ███ ███ ███ ███ 
 MB  └────────────────────────────────────────
      a53d263                             PR

Notable changes:

+1.6 MB at 98e058f

linux/arm64

Binary	PR	vs `main` (`5d99671`)	vs `v1.300070.0`
amazon-cloudwatch-agent	212.2 MB	+0 B	${\color{red}▲}$ +720.9 KB (+0.3%)
amazon-cloudwatch-agent-config-wizard	2.0 MB	+0 B	${\color{green}▼}$ 22 B (-0.0%)
config-downloader	2.0 MB	+0 B	${\color{green}▼}$ 22 B (-0.0%)
config-translator	2.0 MB	+0 B	${\color{green}▼}$ 22 B (-0.0%)
start-amazon-cloudwatch-agent	2.4 MB	+0 B	${\color{red}▲}$ +65.5 KB (+2.8%)
workload-discovery	2.9 MB	+0 B	${\color{red}▲}$ +65.5 KB (+2.3%)
Total	223.5 MB	+0 B	${\color{red}▲}$ +851.8 KB (+0.4%)

windows/amd64

Binary	PR	vs `main` (`5d99671`)	vs `v1.300070.0`
amazon-cloudwatch-agent.exe	236.3 MB	${\color{red}▲}$ +7.7 KB (+0.0%)	${\color{red}▲}$ +2.1 MB (+0.9%)
amazon-cloudwatch-agent-config-wizard.exe	2.1 MB	+0 B	${\color{red}▲}$ +130.6 KB (+6.5%)
config-downloader.exe	2.1 MB	+0 B	${\color{red}▲}$ +130.0 KB (+6.5%)
config-translator.exe	2.1 MB	+0 B	${\color{red}▲}$ +130.6 KB (+6.5%)
start-amazon-cloudwatch-agent.exe	2.3 MB	+0 B	${\color{red}▲}$ +112.1 KB (+5.2%)
workload-discovery.exe	2.9 MB	+0 B	${\color{red}▲}$ +91.6 KB (+3.3%)
Total	247.9 MB	${\color{red}▲}$ +7.7 KB (+0.0%)	${\color{red}▲}$ +2.7 MB (+1.1%)

Investigating size changes

Use go-size-analyzer to compare binaries:

GOEXPERIMENT=jsonv2 go install github.com/Zxilly/go-size-analyzer/cmd/gsa@latest
gsa diff --old <baseline-binary> --new <new-binary>

jefchien · 2026-07-17T19:04:12Z

+    transform/set_cluster_name:
+        error_mode: ignore
+        flatten_data: false
+        log_statements:
+            - context: resource
+              error_mode: ignore
+              statements:
+                - set(resource.attributes["k8s.cluster.name"], "TestCluster")
+        metric_statements: []
+        trace_statements: []


Can you add WithTraceResourceStatements or a WithResourceStatements to that transformprocessor translator so it's consistent?

jefchien · 2026-07-17T21:39:46Z

    transform/set_cluster_name:
        error_mode: ignore
        flatten_data: false
        log_statements: []
-        metric_statements:
+        metric_statements: []
+        trace_statements:
            - context: resource
              error_mode: ignore
              statements:
                - set(resource.attributes["k8s.cluster.name"], "TestCluster")


This isn't quite right. I would expect to see the same statement in each of the *_statements sections.

transform/set_cluster_name: error_mode: ignore flatten_data: false log_statements: - context: resource error_mode: ignore statements: - set(resource.attributes["k8s.cluster.name"], "TestCluster") metric_statements: - context: resource error_mode: ignore statements: - set(resource.attributes["k8s.cluster.name"], "TestCluster") trace_statements: - context: resource error_mode: ignore statements: - set(resource.attributes["k8s.cluster.name"], "TestCluster")

jefchien · 2026-07-17T21:40:48Z

      "type": "object",
      "properties": {
+        "cluster_name": {
+          "description": "The name of the EKS/AKS cluster. Applied to all components under opentelemetry.",


nit:

Suggested change

"description": "The name of the EKS/AKS cluster. Applied to all components under opentelemetry.",

"description": "The name of the k8s cluster. Applied to all telemetry collected under opentelemetry.",

jefchien · 2026-07-17T21:44:56Z

+		stmt := fmt.Sprintf(`set(resource.attributes["k8s.cluster.name"], "%s")`, clusterName)
+		processors.Set(transformprocessor.NewTranslatorWithName("set_cluster_name",
+			transformprocessor.WithTraceResourceStatements([]string{stmt}),
+		))


Each of the set_cluster_name transform processors needs to have all three ResourceStatements set because the same transform processor is used in all pipelines.

Suggested change

))

stmt := fmt.Sprintf(`set(resource.attributes["k8s.cluster.name"], "%s")`, clusterName)

processors.Set(transformprocessor.NewTranslatorWithName("set_cluster_name",

transformprocessor.WithMetricResourceStatements([]string{stmt}),

transformprocessor.WithLogResourceStatements([]string{stmt}),

transformprocessor.WithTraceResourceStatements([]string{stmt}),

))

jefchien · 2026-07-17T21:46:41Z

nit: Can revert the changes for this file.

jefchien · 2026-07-17T21:48:11Z

-// clusterNameRegex restricts cluster_name to safe characters, preventing
-// OTTL injection and template metacharacter issues in YAML templates.
-var clusterNameRegex = regexp.MustCompile(`^[a-zA-Z0-9._-]+$`)
+var otelClusterNameKey = common.ConfigKey(common.OpenTelemetryKey, common.ClusterNameKey)


Define this in common and reuse it for all of the cluster name checks.

jefchien · 2026-07-17T21:48:45Z

 	yamlConfig, err := mapstructure.Marshal(cfg)
 	require.NoError(t, err)
 	yamlStr := toyamlconfig.ToYamlConfig(yamlConfig)
+	// os.WriteFile("./sampleConfig/opentelemetry/container_insights_config.yaml", []byte(yamlStr), 0644)


nit: Don't leave commented out code unnecessarily.

Paamicky added the ready for testing Indicates this PR is ready for integration tests to run label Jul 9, 2026

Paamicky marked this pull request as ready for review July 9, 2026 16:17

Paamicky requested a review from a team as a code owner July 9, 2026 16:17

Paamicky requested review from jefchien, mitali-salvi and sky333999 July 9, 2026 16:18

Paamicky changed the title ~~Add root-level cluster_name and standardize collection_interval for …~~ Add root-level cluster_name and standardize collection_interval for OpenTelemetry features Jul 9, 2026

Paamicky removed the request for review from a team July 10, 2026 16:54

jefchien reviewed Jul 10, 2026

View reviewed changes

Paamicky force-pushed the zeus-V2-json-configs branch 2 times, most recently from 23cab82 to 59cea26 Compare July 10, 2026 20:25

mitali-salvi previously approved these changes Jul 13, 2026

View reviewed changes

Paamicky dismissed mitali-salvi’s stale review via 108055a July 13, 2026 18:53

jefchien reviewed Jul 13, 2026

View reviewed changes

Paamicky force-pushed the zeus-V2-json-configs branch from d7d65ea to 23ade6b Compare July 14, 2026 20:43

mitali-salvi reviewed Jul 16, 2026

View reviewed changes

Paamicky added 7 commits July 16, 2026 18:48

Add root-level cluster_name and standardize collection_interval for …

9cd81e7

…OpenTelemetry features

Add root-level cluster_name and standardize collection_interval for …

1739f5c

…OpenTelemetry features

fix: make cluster_name for prometheus optional

f364e77

fix host metrics azure and aks config

a1b4d9c

fix resource statements for logs and metrics

dc1c85d

change container insights mode to role

1487832

Address comments

67048a9

Paamicky force-pushed the zeus-V2-json-configs branch from a522694 to 939e783 Compare July 16, 2026 19:04

address comments

65265df

Paamicky force-pushed the zeus-V2-json-configs branch from 5689959 to 65265df Compare July 16, 2026 19:12

Merge branch 'main' into zeus-V2-json-configs

5f77c4d

mitali-salvi previously approved these changes Jul 17, 2026

View reviewed changes

Merge branch 'main' into zeus-V2-json-configs

b4658de

Merge branch 'main' into zeus-V2-json-configs

0fd862e

jefchien reviewed Jul 17, 2026

View reviewed changes

Paamicky dismissed mitali-salvi’s stale review via ecdd374 July 17, 2026 21:32

Address comments

8ca11d7

Paamicky force-pushed the zeus-V2-json-configs branch from 10d20a4 to 8ca11d7 Compare July 17, 2026 21:35

Merge branch 'main' into zeus-V2-json-configs

fcafb73

jefchien reviewed Jul 17, 2026

View reviewed changes

	"description": "The name of the EKS/AKS cluster. Applied to all components under opentelemetry.",
	"description": "The name of the k8s cluster. Applied to all telemetry collected under opentelemetry.",

Uh oh!

Conversation

Paamicky commented Jul 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Description of the issue

Description of changes

Tests

License

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

github-actions Bot commented Jul 17, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Binary Size Report

linux/amd64

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Paamicky commented Jul 9, 2026 •

edited

Loading

github-actions Bot commented Jul 17, 2026 •

edited

Loading