Skip to content

Releases: apache/pulsar-helm-chart

pulsar-4.3.0

02 Oct 02:33

Choose a tag to compare

What's Changed

  • Upgrade Oxia image to oxia/oxia:0.14.4 by @lhotari in #625
  • Replace G1GC with generational ZGC and simplify PULSAR_GC options in values.yaml by @lhotari in #626
  • Test with k8s v1.34.0 by @lhotari in #628
  • Add support for ClusterIssuer for certificates by @lhotari in #630
  • Fix tolerations for bookie init by @lhotari in #632
  • Configure automatic compaction of the Bookie entry location index RocksDB database by @lhotari in #633
  • Correct VM Operator CRDs version by @mnit016 in #634
  • Upgrade to Pulsar 4.0.7 by @lhotari in #635

Full Changelog: pulsar-4.2.0...pulsar-4.3.0

pulsar-4.3.0-candidate-1

27 Sep 12:43

Choose a tag to compare

Pre-release

What's Changed

  • Upgrade Oxia image to oxia/oxia:0.14.4 by @lhotari in #625
  • Replace G1GC with generational ZGC and simplify PULSAR_GC options in values.yaml by @lhotari in #626
  • Test with k8s v1.34.0 by @lhotari in #628
  • Add support for ClusterIssuer for certificates by @lhotari in #630
  • Fix tolerations for bookie init by @lhotari in #632
  • Configure automatic compaction of the Bookie entry location index RocksDB database by @lhotari in #633
  • Correct VM Operator CRDs version by @mnit016 in #634
  • Upgrade to Pulsar 4.0.7 by @lhotari in #635

Full Changelog: pulsar-4.2.0...pulsar-4.3.0-candidate-1

pulsar-4.2.0

06 Aug 08:10

Choose a tag to compare

Upgrading from previous releases

TLS configuration for ZooKeeper has changed. Please see detailed instructions for upgrading.

What's Changed

Full Changelog: pulsar-4.1.0...pulsar-4.2.0

pulsar-4.2.0-candidate-1

01 Aug 07:56

Choose a tag to compare

Pre-release

Upgrading from previous releases

TLS configuration for ZooKeeper has changed. Please see detailed instructions for upgrading.

What's Changed

Full Changelog: pulsar-4.1.0...pulsar-4.2.0-candidate-1

pulsar-4.1.0

27 May 20:23
pulsar-4.1.0

Choose a tag to compare

Upgrading to 4.1.0

This version introduces OpenID authentication. Setting auth.authentication.provider is no longer supported, you need to enable the provider with auth.authentication.<provider>.enabled.

What's Changed

New Contributors

Full Changelog: pulsar-4.0.1...pulsar-4.1.0

pulsar-4.1.0-candidate-1

23 May 13:54

Choose a tag to compare

Pre-release

Upgrading to 4.1.0

This version introduces OpenID authentication. Setting auth.authentication.provider is no longer supported, you need to enable the provider with auth.authentication.<provider>.enabled.

What's Changed

New Contributors

Full Changelog: pulsar-4.0.1...pulsar-4.1.0-candidate-1

pulsar-4.0.1

02 May 12:31

Choose a tag to compare

What's Changed

New Contributors

Full Changelog: pulsar-4.0.0...pulsar-4.0.1

pulsar-4.0.1-candidate-1

15 Apr 08:07

Choose a tag to compare

Pre-release

What's Changed

New Contributors

Full Changelog: pulsar-4.0.0...pulsar-4.0.1-candidate-1

pulsar-4.0.0

14 Mar 07:53
pulsar-4.0.0
637cf11

Choose a tag to compare

Apache Pulsar Helm Chart 4.0.0 Release Notes

Breaking Changes

  • Pulsar Proxy Service Configuration Change: Default service type changed from LoadBalancer to ClusterIP for more secure defaults.

  • Monitoring Stack Replacement: Replaced kube-prometheus-stack with victoria-metrics-k8s-stack

    • Victoria Metrics provides full Prometheus compatibility under Apache 2.0 license
    • Required Action: Follow upgrade instructions when migrating from 3.x.x to 4.0.0

Enhanced Monitoring Experience: Usable Metrics and Grafana Dashboards pre-installed

  • Pre-configured and thoroughly tested dashboards now included
  • Dashboards maintained in dedicated repository: https://github.com/lhotari/pulsar-grafana-dashboards
  • Sourced from various Apache 2.0 licensed projects, then optimized and updated for Pulsar 4.0.0 compatibility
  • Resolves long-standing dashboard compatibility issues with recent Pulsar and Bookkeeper
    • Many metrics names have changed since Pulsar 2.10 and previously available dashboards were broken
  • While testing, some bugs were faced in Prometheus 3.2.1 were metrics scraping failed without proper errors even with debug log level and that triggered a switch from kube-prometheus-stack to victoriametrics-k8s-stack and VictoriaMetrics.
    • VictoriaMetrics is more efficient than Prometheus and it doesn't have issues parsing Pulsar metrics.

What's Changed

  • Fix OOM issue on broker wait-zookeeper-ready initContainer by @got-root in #568
  • Increase defaults for ensemble size, write quorum, and ack quorum to 2 by @pdolif in #570
  • Fix pulsar-cluster-initialize / pulsar-init rendering with kustomize by @lhotari in #572
  • [fix] Fix broker service annotations issue, add annotations support to autorecovery service and make handling consistent by @lhotari in #574
  • Upgrade to Pulsar 4.0.3 by @lhotari in #575
  • Upgrade to kube-prometheus-stack 69.x including prometheus-operator 0.80.0 defaulting to Prometheus 3.x by @lhotari in #578
  • Workaround issue with Prometheus 3.0 and metrics by @lhotari in #577
  • Improve kube-prometheus-stack config in values.yaml by adding missing key and some basic comments by @lhotari in #579
  • Restore support for dbStorage_rocksDB_* settings defined in bookkeeper.configData by @lhotari in #580
  • Upgrade default cert-manager version from 1.12.13 to 1.12.16 by @lhotari in #581
  • Replace ">" with "|" to avoid Go Yaml issue go-yaml/yaml#789 by @lhotari in #582
  • Revisit solution to configure Bookkeeper RocksDB settings using individual RocksDB config files by @lhotari in #583
  • Remove the dependency to pulsarctl when generating JWT tokens by @lhotari in #584
  • Use PIP-45 metadata store config to replace deprecated ZK config and make PulsarMetadataBookieDriver configurable in BK by @lhotari in #576
  • Use BookKeeper BP-29 metadataServiceUri to configure bookie metadata store, also when using Zookeeper by @lhotari in #585
  • Add Broker Cache and Sockets dashboards by @lhotari in #586
  • Test with k8s 1.32.2 and upgrade tool versions used in CI by @lhotari in #587
  • Change Pulsar Proxy service load balancer type to ClusterIP by @lhotari in #588
  • Use bookkeeperMetadataServiceUri in broker and make PulsarMetadataClientDriver configurable by @lhotari in #589
  • Remove Prometheus scrape annotations when podmonitors are enabled by @lhotari in #590
  • Remove PodSecurityPolicy support which was removed in Kubernetes 1.25 and bump minimum version to 1.25 by @lhotari in #591
  • Migrate from kube-prometheus-metrics to victoria-metrics-k8s-stack by @lhotari in #592
  • Fix Grafana dashboards for Broker with honorLabels, remove unnecessary *_created metrics and improve docs by @lhotari in #593

New Contributors

Full Changelog: pulsar-3.9.0...pulsar-4.0.0

pulsar-4.0.0-candidate-3

11 Mar 00:13
637cf11

Choose a tag to compare

Pre-release

Apache Pulsar Helm Chart 4.0.0 Release Notes

Breaking Changes

  • Pulsar Proxy Service Configuration Change: Default service type changed from LoadBalancer to ClusterIP for more secure defaults.

  • Monitoring Stack Replacement: Replaced kube-prometheus-stack with victoria-metrics-k8s-stack

    • Victoria Metrics provides full Prometheus compatibility under Apache 2.0 license
    • Required Action: Follow upgrade instructions when migrating from 3.x.x to 4.0.0

Enhanced Monitoring Experience: Usable Metrics and Grafana Dashboards pre-installed

  • Pre-configured and thoroughly tested dashboards now included
  • Dashboards maintained in dedicated repository: https://github.com/lhotari/pulsar-grafana-dashboards
  • Sourced from various Apache 2.0 licensed projects, then optimized and updated for Pulsar 4.0.0 compatibility
  • Resolves long-standing dashboard compatibility issues with recent Pulsar and Bookkeeper
    • Many metrics names have changed since Pulsar 2.10 and previously available dashboards were broken
  • While testing, some bugs were faced in Prometheus 3.2.1 were metrics scraping failed without proper errors even with debug log level and that triggered a switch from kube-prometheus-stack to victoriametrics-k8s-stack and VictoriaMetrics.
    • VictoriaMetrics is more efficient than Prometheus and it doesn't have issues with Pulsar metrics.

What's Changed

  • Fix OOM issue on broker wait-zookeeper-ready initContainer by @got-root in #568
  • Increase defaults for ensemble size, write quorum, and ack quorum to 2 by @pdolif in #570
  • Fix pulsar-cluster-initialize / pulsar-init rendering with kustomize by @lhotari in #572
  • [fix] Fix broker service annotations issue, add annotations support to autorecovery service and make handling consistent by @lhotari in #574
  • Upgrade to Pulsar 4.0.3 by @lhotari in #575
  • Upgrade to kube-prometheus-stack 69.x including prometheus-operator 0.80.0 defaulting to Prometheus 3.x by @lhotari in #578
  • Workaround issue with Prometheus 3.0 and metrics by @lhotari in #577
  • Improve kube-prometheus-stack config in values.yaml by adding missing key and some basic comments by @lhotari in #579
  • Restore support for dbStorage_rocksDB_* settings defined in bookkeeper.configData by @lhotari in #580
  • Upgrade default cert-manager version from 1.12.13 to 1.12.16 by @lhotari in #581
  • Replace ">" with "|" to avoid Go Yaml issue go-yaml/yaml#789 by @lhotari in #582
  • Revisit solution to configure Bookkeeper RocksDB settings using individual RocksDB config files by @lhotari in #583
  • Remove the dependency to pulsarctl when generating JWT tokens by @lhotari in #584
  • Use PIP-45 metadata store config to replace deprecated ZK config and make PulsarMetadataBookieDriver configurable in BK by @lhotari in #576
  • Use BookKeeper BP-29 metadataServiceUri to configure bookie metadata store, also when using Zookeeper by @lhotari in #585
  • Add Broker Cache and Sockets dashboards by @lhotari in #586
  • Test with k8s 1.32.2 and upgrade tool versions used in CI by @lhotari in #587
  • Change Pulsar Proxy service load balancer type to ClusterIP by @lhotari in #588
  • Use bookkeeperMetadataServiceUri in broker and make PulsarMetadataClientDriver configurable by @lhotari in #589
  • Remove Prometheus scrape annotations when podmonitors are enabled by @lhotari in #590
  • Remove PodSecurityPolicy support which was removed in Kubernetes 1.25 and bump minimum version to 1.25 by @lhotari in #591
  • Migrate from kube-prometheus-metrics to victoria-metrics-k8s-stack by @lhotari in #592
  • Fix Grafana dashboards for Broker with honorLabels, remove unnecessary *_created metrics and improve docs by @lhotari in #593

New Contributors

Full Changelog: pulsar-3.9.0...pulsar-4.0.0-candidate-3