For AI agents: A markdown version of this page is available at https://docs.datadoghq.com/byoc-logs/release_notes.md. A documentation index is available at /llms.txt.

BYOC Logs Release Notes

This product is not supported for your selected Datadog site. ().

Overview

This page tracks releases of the BYOC (Bring Your Own Cloud) Logs binary, distributed as a Docker image and bundled by the datadog/cloudprem Helm chart. New features and fixes ship in the binary; the chart packages them for deployment.

Check your installed binary version

Look at the image field on a BYOC Logs pod:

kubectl get pods -n <BYOC_LOGS_NAMESPACE> \
  -o jsonpath='{range .items[*]}{.spec.containers[*].image}{"\n"}{end}' \
  | sort -u

The image tag (for example, :v0.1.26) is the binary version. To see which binary version a Helm chart bundles, run:

helm show chart datadog/cloudprem --version <CHART_VERSION> | grep appVersion

Upgrade

Binary upgrades ship through the Helm chart. See Install BYOC Logs for the chart upgrade command for your platform.

Releases

v0.1.34 — 2026-09-23

Bundled in chart: 0.5.5.

Changed

  • Speeds up indexer decommissioning during rolling updates and scale-downs.
  • Prioritizes monitor queries over other searches, and caches query filters so monitors run faster over shifting time windows.
  • Speeds up field:* existence queries by up to 10x and improves performance for term and histogram aggregations.
  • Adds support for numeric histogram group-by queries.
  • Improves Google Cloud Storage reliability with request timeouts, retries, and credential caching.
  • Removes the embedded web UIs.

Helm chart changes

  • Enables availability zone-aware scheduling by default, and populates the availability zone from pod labels on all Kubernetes versions.
  • Sets QW_ENABLE_IN_MEMORY_INDEXING=true on indexers by default, and sets QW_INDEXING_MAX_WRITE_THROUGHPUT based on indexer.podSize unless indexer.resources is set. Values in indexer.extraEnv or environment take precedence.
  • Fixes the load-estimation configuration in chart 0.5.5: sets QW_DISABLE_LOAD_ESTIMATION=true on searchers by default and removes it from indexer defaults. This disables search load estimation to improve search latency under load; setting the variable on indexers had no effect. Explicit values in searcher.extraEnv or environment take precedence.
  • Uses the readiness endpoint for startup probes, removes default readiness probes, and sets liveness probe timeouts to 5 seconds.
  • Increases the compactor termination grace period to 300 seconds.

v0.1.33 — 2026-08-18

Bundled in chart: 0.5.2. Validated with Observability Pipelines Worker: 2.20.x.

Changed

  • Adds document clustering to group similar logs together and reduce storage footprint by 10% to 20%. To disable document clustering, set QW_DISABLE_DOCS_CLUSTERING=true.
  • Adds support for flat attribute group-by queries.
  • Adds operational metrics for system resource usage, decommissioning, S3 PUT failures, WAL usage, metastore capacity, and split-search outcomes.

Helm chart changes

  • Breaking: Removes the medium pod size. indexer.podSize and searcher.podSize accept large, xlarge, 2xlarge, 4xlarge, 6xlarge, and 8xlarge.
  • Adjusts pod-size CPU and memory requests and limits to account for node reservations and add-ons. This rescales caches, ingest queues, and concurrent split searches accordingly.
  • Enables document clustering by default with config.docs_clustering.
  • Sets indexer and standalone-compactor decommission timeouts to 90% of each workload’s terminationGracePeriodSeconds.
  • Adds a default metastore PodDisruptionBudget and global DNS ndots: 1 setting.
  • Lowers the indexer HPA CPU target to 70% and removes the scale-up stabilization window so indexers scale out under load.

v0.1.32 — 2026-07-21

Bundled in chart: 0.4.6. Validated with Observability Pipelines Worker: 2.20.0 (datadog/observability-pipelines-worker Helm chart 2.20.0).

Changed

  • Adds opt-in PostgreSQL metastore read replica support for search and analytics read paths.
  • Adds an opt-in standalone compactor service to run merge work outside indexer nodes.
  • Reduces S3 DNS lookup churn by caching DNS resolution for S3 clients.
  • Improves control plane stability after actor restarts and metastore overload responses.

Helm chart changes

  • Adds metastore_ro values to deploy a read-only metastore replica pool for scaling metastore reads independently of the writer.
  • Adds enableStandaloneCompactors to run compaction on dedicated workers instead of indexer nodes.
  • Disables ingest v1 by default with QW_DISABLE_INGEST_V1=true; override it with environment.
  • Routes BYOC service traces to the Datadog telemetry intake when datadog.byocTelemetry.enabled is enabled.
  • Uses the dedicated health port for liveness and startup probes.

v0.1.31 — 2026-07-08

Bundled in chart: 0.4.5.

Changed

  • Fixes single-token phrase prefix queries on raw fields so match_phrase_prefix searches return all matching prefix terms instead of being capped by max_expansions.
  • Up to 3x faster intersection for selective terms queries with time range.

Helm chart changes

  • Adds indexer.volumeAttributesClass and searcher.volumeAttributesClass values to provision Kubernetes VolumeAttributesClass resources for indexer and searcher persistent volumes. Use these values to tune volume attributes such as IOPS and throughput. This feature is disabled by default, requires Kubernetes 1.31 or later, and requires driverName when enabled.
  • Fixes the Kubernetes advertise address by setting KUBERNETES_POD_IP from the pod IP instead of the pod name.
  • Disables serviceAccount.automountServiceAccountToken by default to reduce token exposure on pods that do not need Kubernetes API access.
  • Enables securityContext.readOnlyRootFilesystem by default across workloads for defense-in-depth hardening.

v0.1.30 — 2026-06-30

Bundled in chart: 0.4.3.

Changed

  • Reduces search CPU time for nested date histogram queries by up to 20%, with the largest gains on seven-day windows.
  • Adds a dedicated health check listener on port 7284 for CloudPrem component liveness and readiness checks.

Helm chart changes

  • Adds global volumes and volumeMounts values that apply to all CloudPrem components and merge with existing per-component extraVolumes and extraVolumeMounts.
  • Adds global topologySpreadConstraints support, merged with per-component constraints, to spread CloudPrem workload pods across topology domains.
  • Updates CloudPrem services and AWS ALB internal ingress health checks to use the dedicated health endpoint.

v0.1.29 — 2026-06-05

Bundled in chart: 0.4.2.

Changed

  • Faster execution for common log analysis queries, including 2x faster range queries, 1.6x faster cardinality aggregations, and up to 6x faster intersections with range queries.
  • Treats field:* filters as existence queries, and fixes sorting by percentile aggregations.
  • Reduced memory usage for Google Cloud Storage uploads to improve indexing stability.

Helm chart changes

  • Enables BYOC service telemetry by default with datadog.byocTelemetry.enabled; this exports BYOC service logs and metrics only, not customer-ingested logs, metrics, or traces.
  • Deprecates and ignores cloudprem.index.retention, and no longer sets CP_RETENTION_PERIOD.

v0.1.26 — 2026-05-05

Bundled in chart: 0.4.0.

Changed

  • Up to 4x faster term aggregations with order by sub aggregation and up to 1.5x faster cardinality aggregations.

Further reading

Additional helpful documentation, links, and articles:

Plan BYOC Logs updatesDOCUMENTATION more more Install BYOC LogsDOCUMENTATION more more Troubleshoot BYOC LogsDOCUMENTATION more more