Kubernetes Collection v6.0.0 - Full List of Changes

Sourceless mode (enabled by default)

• sumologic.sourcelessMode defaults to true in v6 (was false in v5).
• New flag sumologic.sourcelessModeAck must be set to true to confirm you have read the migration guide; upgrade is blocked until this is set.
• The _source metadata field is no longer populated, so you cannot use _source in search queries. _collector is preserved.
• All Sumo Logic exporters without an explicit endpoint upload data using the installation token through the sourceless path.
• Custom exporters with an explicit endpoint in config.merge continue to use their configured source URL, enabling incremental migration.
• Hosted Collector and its default sources are not deleted automatically on upgrade; use sumologic.cleanupHostedCollector: true to clean up (permanent and irreversible; verify no custom sources exist first).
• sourceType: http is incompatible with sourcelessMode: true. Switch to sourceType: otlp or use an explicit endpoint via config.merge.
• Collection pods register as OpenTelemetry collectors and appear in Manage Data > Collection > OpenTelemetry Collection; filter by cluster=<your-cluster-name> to view pods for a specific cluster.
• GitOps / ArgoCD users with setupEnabled: false must supply sumologic.installationToken manually (token is not created automatically when Terraform/setup job is disabled).

Metrics pipeline unification (enabled by default)

• sumologic.metrics.collector.otelcol.singleLayerPipeline.enabled defaults to true in v6 (was false in v5).
• New flag sumologic.metrics.collector.otelcol.singleLayerPipeline.migrationDocAcknowledged must be set to true regardless of whether the single-layer pipeline is enabled or disabled; upgrade is blocked until this is set. See How to upgrade for detailed migration steps.
• sumologic-metadata-metrics StatefulSet, HPA, Services, and PDB are removed.
• Metrics scraping and metadata enrichment are merged into a single OTel collector pod using two logical pipelines connected by a forward connector.
• Collector resource sizing: memory = collector memory limit + metadata memory limit (apply 1.5x safety multiplier), cpu = collector CPU limit + (total metadata CPU usage / number of collector replicas). With single-layer metrics pipeline, the default metrics collector resources are increased to CPU 768m and memory 768Mi. If your environment requires more resources, you can reconfigure them using sumologic.metrics.collector.otelcol.resources.
• Configuration keys under metadata.metrics.statefulset.* and metadata.metrics.autoscaling.* must be migrated to their sumologic.metrics.collector.otelcol.* equivalents (see How to upgrade for the full mapping table).
• metadata.metrics.config.override is incompatible with single-layer pipeline; use metadata.metrics.config.merge instead.
• Collector pipeline renamed from metrics to metrics/collector; the enrichment pipeline keeps the name metrics. Update any sumologic.metrics.collector.otelcol.config.merge that targets the scraping pipeline to use metrics/collector.
• metadata.metrics.config.merge targeting service.pipelines.metrics continues to work unchanged.
• Prometheus remote write URL hostname changes from <release>-sumologic-metadata-metrics to <release>-sumologic-metrics-collector (same port 9888, same path).
• PVCs from the previous pipeline mode are not automatically deleted when switching between modes and must be manually cleaned up.
• To disable single-layer and restore the 2-layer pipeline, set singleLayerPipeline.enabled: false and run helm upgrade.