hi! When I have replicated subs enabled with geo-replication and auto cluster failover between two clusters, as messages are being sent and consumed the msgInCounter topic stat is always 4x the amount of msgOutCounter. msgOutCounter always equals the actual number of messages sent. Just wondering why msgIn is always so high compared to msgOut with with rep subs? Thanks!

We've just released Apache Pulsar Helm Chart 4.6.1 🎉 The official source release, as well as the binary Helm Chart release, are available at https://www.apache.org/dyn/closer.lua/pulsar/helm-chart/4.6.1/?action=download The helm chart index at https://pulsar.apache.org/charts/ has been updated and the release is also available directly via helm. Release Notes: https://github.com/apache/pulsar-helm-chart/releases/tag/pulsar-4.6.1 Docs: https://github.com/apache/pulsar-helm-chart#readme and https://pulsar.apache.org/docs/helm-overview ArtifactHub: https://artifacthub.io/packages/helm/apache/pulsar/4.6.1 Thanks to all the contributors who made this possible.

🎉 <!channel> Apache Pulsar 5.0.0-M1 is out! This is a milestone preview of the upcoming Pulsar 5.0 LTS, released early to gather feedback. ⚠️ Not for production — breaking changes may still land before the final 5.0 LTS. What's new in 5.0, previewed here: 🔄 Scalable Topics — a new

topic://

topic type that sizes itself to its load. The broker splits hot key-range segments and merges cold ones at runtime, no downtime, no broken per-key ordering, and no partition count to pick up front. ✨ New V5 Java client (

pulsar-client-v5

) — a clean-slate redesign with three purpose-built consumers (Stream, Queue, Checkpoint). Works with your existing topics, a consumer can subscribe to a whole namespace, and it's the only way to use Scalable Topics. The classic client API stays fully supported. 🛢️ Oxia is now the recommended metadata store for new clusters, with live zero-downtime migration from ZooKeeper. ZooKeeper is still fully supported; the etcd backend is removed in 5.0. Also notable: structured (slog-style) logging, IO connectors in a dedicated repo, the

javax.*

→

jakarta.*

migration, and a move from Maven to Gradle. 📚 Full announcement: https://pulsar.apache.org/blog/2026/06/23/announcing-apache-pulsar-5-0-m1/ ⬇️ Download: https://pulsar.apache.org/download 📝 Release notes: https://pulsar.apache.org/release-notes/versioned/pulsar-5.0.0-M1/ Thanks to everyone who made this release possible! pulsarlogo

Pulsar go client from the stream native repo support for the AVRO and JSON message schema .

📣 pulsarlogo [ANNOUNCE] Apache Pulsar 4.0.12 released pulsarlogo 📣 The Apache Pulsar team is proud to announce Apache Pulsar version 4.0.12. Pulsar is a highly scalable, low latency messaging platform running on commodity hardware. It provides simple pub-sub semantics over topics, guaranteed at-least-once delivery of messages, automatic cursor management for subscribers, and cross-datacenter replication. For Pulsar release details and downloads, visit: pulsar.apache.org/download Release Notes are at: pulsar.apache.org/release-notes/versioned/pulsar-4.0.12 We would like to thank the contributors that made the release possible. Regards, The Pulsar Team

📣 pulsarlogo [ANNOUNCE] Apache Pulsar 4.2.3 released pulsarlogo 📣 The Apache Pulsar team is proud to announce Apache Pulsar version 4.2.3. Pulsar is a highly scalable, low latency messaging platform running on commodity hardware. It provides simple pub-sub semantics over topics, guaranteed at-least-once delivery of messages, automatic cursor management for subscribers, and cross-datacenter replication. For Pulsar release details and downloads, visit: pulsar.apache.org/download Release Notes are at: pulsar.apache.org/release-notes/versioned/pulsar-4.2.3 We would like to thank the contributors that made the release possible. Regards, The Pulsar Team

Hi Everyone 👋 , got a question around the apache pulsar helm chart. I noticed the

master

version is now supporting

components.function_worker

(separate functions component instead of embedded functions in the broker). However this has not been published in the latest release 4.6.1 version yet. Are there any plans to release it please? Thanks

We've just released Apache Pulsar Helm Chart 4.7.0 🎉 The official source release, as well as the binary Helm Chart release, are available at apache.org/dyn/closer.lua/…/4.7.0?action=download The helm chart index at pulsar.apache.org/charts has been updated and the release is also available directly via helm. Release Notes: github.com/apache/…/pulsar-4.7.0 Docs: github.com/apache/pulsar-helm-chart#… and pulsar.apache.org/docs/helm-overview ArtifactHub: artifacthub.io/packages/…/4.7.0 Thanks to all the contributors who made this possible.

hi ,In the java pulsar admin we have the resetcursor with exclude the last message id ,is same approach is available in pulsar admin go ? public ResetCursorData(long ledgerId, long entryId, boolean isExcluded) { this.ledgerId = ledgerId; this.entryId = entryId; this.isExcluded = isExcluded; } func (s *subscriptions) ResetCursorToMessageID(topic utils.TopicName, sName string, id utils.MessageID) error { return s.ResetCursorToMessageIDWithContext(context.Background(), topic, sName, id) }

Do you have an Apache Pulsar use case you'd like to share with a wider audience? The final CFP deadline for Data Streaming Summit 2026 - The Data Streaming + Agent Infra Conference is in about 32 hours from now. Please submit your proposals now.

📣 pulsarlogo [ANNOUNCE] Apache Pulsar 4.0.13 released pulsarlogo📣 The Apache Pulsar team is proud to announce Apache Pulsar version 4.0.13. Pulsar is a highly scalable, low latency messaging platform running on commodity hardware. It provides simple pub-sub semantics over topics, guaranteed at-least-once delivery of messages, automatic cursor management for subscribers, and cross-datacenter replication. For Pulsar release details and downloads, visit: pulsar.apache.org/download Release Notes are at: pulsar.apache.org/release-notes/versioned/pulsar-4.0.13/ We would like to thank the contributors that made the release possible. Regards, The Pulsar Team

📣 pulsarlogo [ANNOUNCE] Apache Pulsar 4.2.4 released pulsarlogo📣 The Apache Pulsar team is proud to announce Apache Pulsar version 4.2.4. Pulsar is a highly scalable, low latency messaging platform running on commodity hardware. It provides simple pub-sub semantics over topics, guaranteed at-least-once delivery of messages, automatic cursor management for subscribers, and cross-datacenter replication. For Pulsar release details and downloads, visit: pulsar.apache.org/download Release Notes are at: pulsar.apache.org/release-notes/versioned/pulsar-4.2.4/ We would like to thank the contributors that made the release possible. Regards, The Pulsar Team

Hi, has anyone encountered a transaction pending-ack replay thread getting stuck indefinitely? On Pulsar 4.0.10, I observed a pulsar-transaction-executor-N-1 thread stuck in MLPendingAckStore$PendingAckReplay.run(), repeatedly calling Thread.sleep(1). Because these executors are single-threaded and shared by multiple subscriptions, one stuck replay can prevent consumers from being created on unrelated subscriptions. I also observed broker pod restarts around the same time. I suspect the stuck replay may block the health-check subscription when it hashes to the affected executor, causing Kubernetes health checks to fail and restart the broker pod. My current theory is: - A managed-ledger read is issued, but its callback never arrives. - outstandingReadsRequests remains 1, the entry queue stays empty, and isReadable remains true. - PendingAckReplay therefore loops forever. - Normally, ManagedLedgerImpl.checkReadTimeout() should terminate the read, but this protection is disabled when managedLedgerReadEntryTimeoutSeconds=0. - A possible trigger is an empty or newly rolled-over pending-ack ledger being deleted while replay is reading it, because new ledgers initially have timestamp=0 and system ledgers use retention 0. I found related issues #14852, #23506, and #24401, but none appears to fully address pending-ack replay when the read callback never returns. Has anyone seen this behavior? Does this analysis seem plausible? Could this explain the broker pod restarts, and is there a recommended configuration to avoid it besides setting managedLedgerReadEntryTimeoutSeconds to a non-zero value? I have a repeatable local reproduction and thread dumps that I’m happy to share.

Hi everyone! I am launching a single node Pulsar cluster via docker-compose and following documentation for setting the yaml file up. I would like to create a pre-configured sink and pre-configured tenants/ns/topics. I am doing this just by creatinga. script that has the pulsar-cli commands to do those operations. I would like those resources to be available to the user on broker start up. My question is, where would it be best to pass in this script? as an entrypoint in the broker container? or the pulsar-init container?

another question regarding schemas .. we are streaming apache arrow record batch tables. Currently, we have to serialize the record batch into bytes, producer sends arrow bytes to consumer, consumer receives and acks, and then we have to deserialize it. Is there any ability to natively send arrow record batches without serialization.

@RunLLM If the message size is configured to maxMessageSize=5242880 in broker.conf what is the max entry size in bookie .

Hi! What happens to messages that are sent using a Pulsar producer during a Pulsar broker temporary outage? Are they stored in a producer backlog/queue until the broker is back up? are there settings to configure this? The producer also disconnects/crashes when the broker is unavailable?

📣 pulsarlogo [ANNOUNCE] Apache Pulsar 5.0.0-M2 released pulsarlogo 📣 The Apache Pulsar team is proud to announce Apache Pulsar version 5.0.0-M2. 5.0.0-M2 is a milestone release for the upcoming Pulsar 5.0 LTS release, made to gather user feedback early. It is not meant for production use cases, since breaking changes can be introduced between the milestone releases and the final Pulsar 5.0 LTS release. Pulsar is a highly scalable, low latency messaging platform running on commodity hardware. It provides simple pub-sub semantics over topics, guaranteed at-least-once delivery of messages, automatic cursor management for subscribers, and cross-datacenter replication. For Pulsar release details and downloads, visit: pulsar.apache.org/download Release Notes are at: pulsar.apache.org/release-notes/versioned/pulsar-5.0.0-M2/ We would like to thank the contributors that made the release possible. Regards, The Pulsar Team

Hey @Lari Hotari, do you have any idea if this is something which can be tackled anytime soon? github.com/apache/pulsar/issues/24914 We are still experiencing lot of exceptions cause of this, so if there is anything which I can provide (open a discussion, providing examples etc. ) I am happy to do so.

Pulsar 5.0.0, 4.2.5 and 4.0.14 release candidates are out for voting: https://apache-pulsar.slack.com/archives/C5ZSVEN4E/p1790887582019909 https://apache-pulsar.slack.com/archives/C5ZSVEN4E/p1790895536679729

📣 pulsarlogo [ANNOUNCE] The Apache Pulsar team is proud to announce Apache Pulsar version 4.0.14. pulsarlogo📣 Pulsar is a highly scalable, low latency messaging platform running on commodity hardware. It provides simple pub-sub semantics over topics, guaranteed at-least-once delivery of messages, automatic cursor management for subscribers, and cross-datacenter replication. For Pulsar release details and downloads, visit: pulsar.apache.org/download Release Notes are at: pulsar.apache.org/release-notes/versioned/pulsar-4.0.14/ We would like to thank the contributors that made the release possible. Regards, The Pulsar Team

📣 pulsarlogo [ANNOUNCE] The Apache Pulsar team is proud to announce Apache Pulsar version 4.2.5. pulsarlogo 📣 Pulsar is a highly scalable, low latency messaging platform running on commodity hardware. It provides simple pub-sub semantics over topics, guaranteed at-least-once delivery of messages, automatic cursor management for subscribers, and cross-datacenter replication. For Pulsar release details and downloads, visit: pulsar.apache.org/download Release Notes are at: pulsar.apache.org/release-notes/versioned/pulsar-4.2.5/ We would like to thank the contributors that made the release possible. Regards, The Pulsar Team

<!channel> 📣 pulsarlogo [ANNOUNCE] Apache Pulsar 5.0.0 LTS released pulsarlogo 📣 🎉 The Apache Pulsar team is proud to announce Apache Pulsar version 5.0.0, the new long-term support (LTS) release. Pulsar is a highly scalable, low latency messaging platform running on commodity hardware. It provides simple pub-sub semantics over topics, guaranteed at-least-once delivery of messages, automatic cursor management for subscribers, and cross-datacenter replication. Apache Pulsar 5.0 LTS: Scalable Topics and Faster Existing Workloads Scalable Topics are a new kind of topic that grows and shrinks with demand while preserving per-key ordering, so applications no longer have to choose a partition count up front. They come with a new Java client API (v5), a language-neutral client specification for all SDKs, and Oxia as the recommended metadata store for new clusters. Pulsar 5.0 also brings major performance and reliability improvements for the topics and applications you already run. For example, on topics with a large number of producers publishing small, non-batched messages at a high rate, consumers now keep up with the producers instead of falling behind. Reliability fixes cover geo-replication, subscriptions, delayed delivery, and acknowledgment recovery. You can upgrade a 4.0.x or 4.2.x cluster to 5.0 and keep your existing topics, the existing Java client (now called the v4 client), and ZooKeeper. The improvements apply without application changes, and you can adopt the v5 client API, Scalable Topics, or Oxia later, on your own schedule. Blog post: pulsar.apache.org/blog/…/announcing-apache-pulsar-5-0 Release highlights and upgrading: pulsar.apache.org/docs/5.0.x/release-highlights Upgrading to Pulsar 5.0.x: pulsar.apache.org/docs/5.0.x/administration-upgrade-to-5.0.x Java client dependency configuration: pulsar.apache.org/docs/client-libraries/java-dependency-configuration Downloads: pulsar.apache.org/download Release notes: pulsar.apache.org/release-notes/versioned/pulsar-5.0.0 We would like to thank the contributors that made the release possible, and everyone who tested the 5.0.0-M1 and 5.0.0-M2 milestones and shared feedback. Regards, The Pulsar Team

Hello everyone, a general question on cluster upgrade from very old to new. I am running version 2.10.1 currently and planning to upgrade it to 4 then 5. My security team has highlighted few critical CVE which are fixed in 2.10.6 and I have to upgrade the cluster. Question is, can I directly upgrade cluster from 2.10.1 to 2.10.6 or I have to hop every patch version in between ? Once its on 2.10.6 I will move it to 2.11 then 3 and 4. TIA 🙏

@Lari Hotari @merlimat red.anthropic.com/oss-scanner

<!channel> 🚨 Several security advisories (CVEs) for the Apache Pulsar project were published today. 🚨 For Apache Pulsar components, these are fixed in the 4.0.14, 4.2.5, and 5.0.0 releases: • CVE-2026-105945: Apache Pulsar: Improper Validation of Package Name Components Allows Cross-Tenant Package Access in Filesystem Package Storage • CVE-2026-105942: Apache Pulsar: Original principal of proxied requests authorized with the proxy's authentication data, allowing privilege escalation • CVE-2026-105938: Apache Pulsar: Subscription authorization policies not enforced by namespace-level backlog and unsubscribe operations and by topic expire-all • CVE-2026-105937: Apache Pulsar: Missing authorization of the shadow topic source allows cross-tenant message disclosure and deletion of source topic data • CVE-2026-105936: Apache Pulsar: Functions connectors list endpoint does not apply the super-user check of the worker connectors endpoint • CVE-2026-105935: Apache Pulsar: Function trigger publishes to the input topic without checking the caller's produce permission • CVE-2026-105934: Apache Pulsar: Missing tenant admin authorization check when truncating a partitioned topic • CVE-2026-105933: Apache Pulsar: Namespace encryption requirement not enforced after a failure during topic initialization • CVE-2026-105932: Apache Pulsar: Missing topic policy authorization on non-persistent topic entry filter endpoints • CVE-2026-105931: Apache Pulsar: Missing authorization on transaction admin endpoints exposes transaction coordinator and transaction statistics of other tenants • CVE-2026-105930: Apache Pulsar: Improper Authorization of Function, Source and Sink Package URLs Allows Cross-Tenant Package Download • CVE-2026-105927: Apache Pulsar: Improper authorization of WebSocket consumer dead letter topic allows writes without PRODUCE permission • CVE-2026-105926: Apache Pulsar: Missing Topic Authorization on Binary Protocol GetSchema and GetOrCreateSchema Commands • CVE-2026-105869: Apache Pulsar: MultiRolesTokenAuthorizationProvider uses roles from unverified tokens for HTTP requests that fall back to the anonymous user role In Apache Pulsar Manager, this advisory covers multiple critical vulnerabilities (CVSS 10.0), and no fix is available. The project’s end of life is announced with this CVE: • CVE-2026-105944: Apache Pulsar Manager: Missing authentication and authorization and disclosure of the Pulsar admin credential; end of life Unfortunately, there’s a delay in updating the website. The recent security advisory information will also be available at pulsar.apache.org/security#… once the GitHub Actions workflow has run. (There have been some GitHub availability incidents today.)

interesting google has added a low latency link to google cloud storage

seaweedfs.com/blog/s3-over-rdma interesting developments for object storage