feat(vector): add support for raising file descriptor limits

.github/actions/spelling/patterns.txt

@@ -19,6 +19,9 @@ https?://\S+
 # Version strings
 v\d+\.\d+(\.\d+)?
 
+# AsciiDoc source/listing blocks (---- delimited)
+(?s)^----$.*?^----$
+
 # AsciiDoc directives and macros
 ^include::.*$
 ^ifdef::.*$

docs/features/raise_fd_limit.adoc

@@ -0,0 +1,7 @@
+== Vector Collector: Raising File Descriptor Limits
+
+The operator always sets `VECTOR_RAISE_FD_LIMIT=true` on the collector container so Vector raises its file descriptor soft limit to the hard limit at startup. This prevents "Too many open files" errors when Vector monitors a large number of log files concurrently.
+
+Systems often default to restrictive file descriptor soft limits (e.g. 1024 on Linux), which can cause Vector to fail when processing many log sources simultaneously. Vector reads `VECTOR_RAISE_FD_LIMIT` natively and raises the soft limit without requiring manual sysadmin intervention. The hard limit remains the safety ceiling set by the OS/container runtime.
+
+This behavior is not configurable. A collector tuning option may be added in the future if needed.

docs/reference/operator/api_observability_v1.adoc

@@ -61,6 +61,7 @@ Type:: object
 |networkPolicy|object|  Define the Network Policy for the Collector
 |nodeSelector|object|  Define nodes for scheduling the pods.
 |resources|object|  The resource requirements for the collector
+|terminationGracePeriodSeconds|int|  TerminationGracePeriodSeconds defines the termination grace period for collector pods in seconds. If not specified, the default is 10 seconds.
 |tolerations|array|  Define the tolerations the collector pods will accept
 |======================
 
@@ -651,6 +652,10 @@ Type:: object
 
 Type:: object
 
+=== .spec.collector.terminationGracePeriodSeconds
+
+Type:: int
+
 === .spec.collector.tolerations[]
 
 Type:: array

internal/collector/collector.go

@@ -239,6 +239,7 @@ func (f *Factory) NewCollectorContainer(inputs internalobs.Inputs, outputs inter
 		{Name: "OPENSHIFT_CLUSTER_ID", Value: clusterID},
 		{Name: "POD_IP", ValueFrom: &v1.EnvVarSource{FieldRef: &v1.ObjectFieldSelector{APIVersion: "v1", FieldPath: "status.podIP"}}},
 		{Name: "POD_IPS", ValueFrom: &v1.EnvVarSource{FieldRef: &v1.ObjectFieldSelector{APIVersion: "v1", FieldPath: "status.podIPs"}}},
+		{Name: "VECTOR_RAISE_FD_LIMIT", Value: "true"},
 	}
 	collector.Env = append(collector.Env, utils.GetProxyEnvVars()...)
 

internal/collector/collector_test.go

@@ -143,6 +143,10 @@ var _ = Describe("Factory#Daemonset", func() {
 				Expect(collector.Env).To(IncludeEnvVar(v1.EnvVar{Name: "VECTOR_LOG", Value: logLevelDebug}))
 			})
 
+			It("should set VECTOR_RAISE_FD_LIMIT to true", func() {
+				Expect(collector.Env).To(IncludeEnvVar(v1.EnvVar{Name: "VECTOR_RAISE_FD_LIMIT", Value: "true"}))
+			})
+
 			Context("the volume mounts", func() {
 				It("should mount all output configmaps", func() {
 					Expect(collector.VolumeMounts).To(IncludeVolumeMount(

test/framework/functional/vector/deploy.go

@@ -38,6 +38,7 @@ func (c *VectorCollector) DeployConfigMapForConfig(name, config, clfName, clfYam
 
 func (c *VectorCollector) BuildCollectorContainer(b *runtime.ContainerBuilder, nodeName string) *runtime.ContainerBuilder {
 	return b.AddEnvVar("VECTOR_LOG", common.AdaptLogLevel()).
+		AddEnvVar("VECTOR_RAISE_FD_LIMIT", "true").
 		AddEnvVarFromFieldRef("POD_IP", "status.podIP").
 		AddEnvVar("NODE_NAME", nodeName).
 		AddEnvVar("VECTOR_INTERNAL_LOG_RATE_LIMIT", "0").

test/functional/misc/disk_buffer_corruption_test.go

@@ -313,7 +313,7 @@ func corruptBufferFile(framework *functional.CollectorFunctionalFramework, datFi
 
 	const chunkSize = 65536
 	for i := 0; i < len(corruptedB64); i += chunkSize {
-		end := min(i + chunkSize, len(corruptedB64))
+		end := min(i+chunkSize, len(corruptedB64))
 		chunk := corruptedB64[i:end]
 		appendCmd := fmt.Sprintf("echo -n '%s' >> '%s'", chunk, stagingFile)
 		_, err = framework.RunCommand(functional.ToolsContainerName, "bash", "-c", appendCmd)

test/functional/misc/raise_fd_limit_test.go

@@ -0,0 +1,34 @@
+package misc
+
+import (
+	. "github.com/onsi/ginkgo/v2"
+	. "github.com/onsi/gomega"
+	obs "github.com/openshift/cluster-logging-operator/api/observability/v1"
+	"github.com/openshift/cluster-logging-operator/internal/constants"
+	"github.com/openshift/cluster-logging-operator/test/framework/functional"
+	testruntime "github.com/openshift/cluster-logging-operator/test/runtime/observability"
+)
+
+var _ = Describe("[Functional][Misc][RaiseFdLimit] Vector raise-fd-limit", func() {
+
+	var framework *functional.CollectorFunctionalFramework
+
+	BeforeEach(func() {
+		framework = functional.NewCollectorFunctionalFramework()
+		testruntime.NewClusterLogForwarderBuilder(framework.Forwarder).
+			FromInput(obs.InputTypeInfrastructure).
+			ToHttpOutput()
+	})
+
+	AfterEach(func() {
+		framework.Cleanup()
+	})
+
+	It("should start with VECTOR_RAISE_FD_LIMIT=true", func() {
+		Expect(framework.Deploy()).To(BeNil())
+
+		out, err := framework.RunCommand(constants.CollectorName, "sh", "-c", "echo $VECTOR_RAISE_FD_LIMIT")
+		Expect(err).To(BeNil())
+		Expect(out).To(ContainSubstring("true"))
+	})
+})

test/helpers/types/types.go

@@ -165,7 +165,7 @@ type OpenshiftMeta struct {
 	//+optional
 	Labels map[string]string `json:"labels,omitempty"`
 
-	//Sequence is increasing id used in conjunction with the timestamp to estblish a linear timeline
+	//Sequence is increasing id used in conjunction with the timestamp to establish a linear timeline
 	//of log records.  This was added as a workaround for logstores that do not have nano-second precision.
 	Sequence OptionalInt `json:"sequence,omitempty"`
 }