Emergency Incident Response
Cluster partitioned, publishers blocked, or a node that won't rejoin. With response targets from 15 minutes, by plan, a senior RabbitMQ engineer joins a live bridge with access to diagnose — not a ticket acknowledgement.
Tanzu RabbitMQ support
Last reviewed by Tyler Eastridge
AceMQ supports Tanzu RabbitMQ 24/7 across Americas, EMEA and APAC: the Kubernetes deployment with the Cluster Operator, and the OVA on vSphere. Most calls start with one of three problems: a Cluster Operator reconcile loop that never reaches ready, a quorum queue short of members after a node replacement, or an upgrade blocked by a Broadcom entitlement. Link sentence (placed directly under the opener): Running open-source RabbitMQ rather than the Tanzu distribution? See our RabbitMQ support page.
AceMQ is a Featured Authorized Partner named by RabbitMQ, with direct access to the RabbitMQ core team, a Leading Broadcom Tanzu partner and a Leading Global Services Partner, and Broadcom's 2025 Tanzu Partner of the Year. We support Tanzu RabbitMQ on both the Kubernetes and OVA deployment models — Cluster Operator reconcile loops, quorum queues short of members, Erlang distribution broken by cert rotation, and the entitlement problems that block your upgrade.
Planning a build, upgrade or migration instead? See AceMQ Tanzu RabbitMQ consulting and services.
Customers




Partners


Incident Triage
Reconcile loops, short quorum queues, blocked publishers and stalled upgrades on Tanzu RabbitMQ, with the operator log, event or rabbitmq-queues command we use first.
| Symptom you're seeing | What we check first |
|---|---|
| The Cluster Operator is in a reconcile loop and the cluster never becomes ready | Read the operator logs and the RabbitmqCluster events. Typical causes are a StatefulSet update the operator cannot complete, a PVC that cannot be resized because the storage class does not allow expansion, an image the cluster cannot pull (registry credentials for the Broadcom-distributed image), a resource request the nodes cannot satisfy, or a conflict between a Helm value and an operator-managed field. We fix the cause and then check the cluster for data-safety implications before the operator proceeds. |
| A quorum queue has fewer members than expected after a node replacement | Quorum queues are Raft-based. When a node is replaced, the queue does not move to the new node on its own. rabbitmq-queues quorum_status shows the members; rabbitmq-queues grow adds a member to a new node, and rabbitmq-queues rebalance quorum evens out leaders. A quorum queue with two of three members is one failure from losing writes. We check membership on every queue after any node change, and use rabbitmq-diagnostics check_if_node_is_quorum_critical before taking a node down. |
| Nodes will not rejoin after a certificate rotation | Inter-node communication uses Erlang distribution, which may be TLS in a hardened deployment. A cert-manager rotation that changes the chain or SANs can break node-to-node connections while client connections still work. We check the certificates on both ends, the TLS distribution configuration, and the order of the rolling restart. |
| An upgrade is blocked: the Broadcom portal says you are not entitled | Tanzu RabbitMQ is distributed through the Broadcom Support Portal under your Broadcom contract. Entitlement problems appear as missing downloads, an image you cannot pull, or a version listed for a different product line. We help you resolve it through the right channel, and plan a path that does not leave the cluster stranded while it is sorted out. |
| Memory alarm triggered and publishers blocked while memory looks moderate | The alarm fires when RabbitMQ's memory use crosses vm_memory_high_watermark of the detected memory, which inside a container is the container limit. rabbitmq-diagnostics memory_breakdown shows where it goes: queue processes, binary heap, connections, quorum queue WAL, metadata store. Quorum queues keep recent messages in memory through the Raft log, so large backlogs and large messages weigh more than they did on classic queues. We tune the watermark, WAL and queue settings, and the Kubernetes limit together, since a limit below the watermark's expectation gets the pod OOMKilled first. |
| A rolling upgrade between Tanzu RabbitMQ versions stalls | Feature flags must be enabled in order before an upgrade past certain versions (rabbitmqctl list_feature_flags), the operator and the broker image versions must be compatible, and a quorum queue leader must have a caught-up follower before its node restarts. We check all three before the window, then run the upgrade one pod at a time with rabbitmq-queues checks between pods. |
| Classic mirrored queues misbehave after migration to quorum queues | Mirrored classic queues were removed in the 4.x line, so the migration has a deadline. Behaviour differs: quorum queues need x-queue-type set at declaration, have different delivery limit semantics, do not support some classic features such as priorities in the same way, and need more disk for the Raft log. We inventory queues, test consumers against the new semantics and migrate in groups. |
| Shovel or federation link flapping between sites | We check the link's credentials, prefetch, ack mode and reconnect delay, network between foundations, and whether the Tanzu distributed shovel setting spreads shovels as expected. For disaster recovery with the warm-standby feature, we verify the replication connection and the standby's state. |
| The deployment model | Kubernetes and OVA behave differently. On vSphere OVA, a noisy neighbour or a storage latency spike on a datastore looks like a broker fault. On Kubernetes, node pressure and storage class limits do. We read the platform signals first. |
A P1 ends with a report on the operator event, queue membership, certificate or memory setting at fault and its fix.
Deployment coverage
| Deployment | What AceMQ covers | What Broadcom covers | Excluded |
|---|---|---|---|
| Tanzu RabbitMQ on Kubernetes | Cluster Operator and Messaging Topology Operator behaviour, StatefulSets, storage, quorum queues, streams, TLS, upgrades | Broadcom supports the product and its fixes under your Broadcom contract | A third party cannot ship a patched Tanzu RabbitMQ binary |
| Tanzu RabbitMQ OVA on vSphere | Broker and cluster behaviour, host-level contention diagnosis, upgrade and migration to Kubernetes | Broadcom for the OVA product | vSphere platform faults outside the agreed scope |
| Tanzu Kubernetes Grid and Tanzu Application Service | The RabbitMQ layer on these platforms | Broadcom for the platform | Platform defects |
| OpenShift, EKS, AKS and GKE | The RabbitMQ layer and its operators | The platform vendor | Control-plane faults |
| Warm standby and multi-site replication | Link health, promotion and failback runbooks, rehearsals | Broadcom for the feature as a product | Network provider faults |
| Quorum queues, streams and classic queues | Queue design, membership, leaders, delivery limits, migration from mirrored classic queues | Broadcom for product defects | Application consumer code is reviewed, not rewritten |
| Air-gapped and disconnected foundations | Image mirroring and upgrade rehearsal inside your boundary | Broadcom for the product | Anything needing outbound access |
Better together
Broadcom, through VMware Tanzu, sells Tanzu RabbitMQ with a support entitlement. The product, the fixes and the entitlement belong to Broadcom. AceMQ does not replace that licence; it sits beside it, runs the incident, checks the platform around the broker and takes product defects to Broadcom with the evidence ready.
| Broadcom subscription | AceMQ engineers alongside it | |
|---|---|---|
| What it covers | The Tanzu RabbitMQ product: defects, fixes and the commercial features | The broker as deployed: Cluster Operator, storage, quorum queues, TLS, upgrades, client behaviour and the platform under it |
| Hours and response | Per the terms of your subscription | 24/7 across Americas, EMEA and APAC; response targets from 15 minutes, by plan |
| Open-source RabbitMQ | Not covered by a Tanzu subscription | Covered by the same AceMQ contract |
| Root cause analysis | Per Broadcom's process | Written root cause analysis on every P1 |
| Product defects | Ships fixed builds and handles product defects | We diagnose, gather the evidence and work the escalation with you |
| Price | Set by the Broadcom subscription | $24,000 to $72,000 a year (USD), set by SLA, scope and deployment. Quote reply within 1 hour |
If you need the commercial features, the entitlement or the fixed builds, you buy them from Broadcom. If you also want senior engineers on call across time zones, that is a separate contract that sits beside it. Many estates do both, and AceMQ is a Leading Global Services Partner for Broadcom.
Each guide says what changed, who it affects and what to decide first.
What's Included
The Cluster Operator, Messaging Topology Operator, quorum queues, TLS and upgrades are handled as one deployment, on Kubernetes or on vSphere OVA.
Cluster partitioned, publishers blocked, or a node that won't rejoin. With response targets from 15 minutes, by plan, a senior RabbitMQ engineer joins a live bridge with access to diagnose — not a ticket acknowledgement.
Every P1 closes with a written RCA: what failed, why, the fix applied, and the configuration, manifest, or client change that prevents recurrence. Delivered as standard, not on request.
Direct access to the RabbitMQ core team plus partner-level escalation into Broadcom engineering. When an issue is a genuine defect rather than configuration, it goes upstream through a channel you do not have directly.
Quorum versus classic queue selection, prefetch and publisher-confirm strategy, memory watermarks against real container limits, and disk sizing for the write-ahead log under your actual message profile.
Broadcom portal migrations, split site accounts, subscription conversions, and support contract mapping. We get the download and the entitlement working again instead of handing you a portal link.
Version upgrades with feature-flag sequencing, mirrored to quorum queue migration, OVA to Kubernetes moves, and migration between open-source and Tanzu RabbitMQ in either direction.
Why AceMQ
Direct access to the RabbitMQ core team. When an incident turns out to be a genuine broker defect — and occasionally it does — we escalate to the people who maintain the code rather than filing a ticket and waiting with you.
Broadcom Expert Advantage Tanzu Services Partner of the Year, AMER (North America), 2025. Our engineers hold current Broadcom certification across the Tanzu portfolio, and we escalate through partner channels on product defects and entitlement problems alike.
The same senior engineers stay on your account. They know your cluster size, your queue types, and your client libraries — so a P1 call starts with diagnosis, not twenty minutes of you explaining your topology.
You reach a senior RabbitMQ engineer directly by phone, email, or Slack. No help desk collecting information to pass along, and no escalation approval process standing between you and someone who can actually fix it.
Most Tanzu RabbitMQ incidents on Kubernetes have a platform cause and a broker symptom: a storage class that can't sustain the write-ahead log, a PodDisruptionBudget that permitted the drain that cost you quorum, a NetworkPolicy blocking Erlang distribution. We diagnose both layers.
Round-the-clock coverage for customers in 26+ countries. Your 3am cluster partition is someone's mid-afternoon — no overnight skeleton crew, no waiting for a region to wake up.
Tanzu RabbitMQ on TKG, OpenShift and managed Kubernetes services, OVA appliances on vSphere, warm-standby pairs across sites, and disconnected foundations.
Version coverage
| Tanzu RabbitMQ series | Supported by AceMQ |
|---|---|
| 4.4 | Yes |
| 4.3 | Yes |
| 4.2 | Yes |
| 4.1 | Yes |
| 4.0 | Yes |
| 3.13 | Yes |
| Earlier series | Ask at scoping |
Dated sources: Broadcom product release advisory for VMware Tanzu RabbitMQ on Kubernetes 4.3.0, 4.2.6, 4.1.11, 4.0.20 and 3.13.15 (support.broadcom.com, dated 8 May 2026 per the search listing); rabbitmq.com/release-information, read 17 November 2026, which lists 4.4 as first released 20 October 2026. This page does not publish lifecycle dates: Broadcom's support portal is the authority for them.
In practice
Tanzu RabbitMQ is the commercial distribution of RabbitMQ from Broadcom. The RabbitMQ commercial features page (blog.rabbitmq.com/commercial-features, read 17 November 2026) lists what the commercial edition adds. Each feature changes what a support engineer must check.
Source: RabbitMQ blog, "Commercial Features", blog.rabbitmq.com/commercial-features, read 17 November 2026. The page does not use the term "Warm Standby Replication"; use Broadcom's own name once checked.
| Tanzu RabbitMQ feature (as listed by RabbitMQ) | What it means in support |
|---|---|
| Support from core engineers, 24x7, from the team that builds and maintains RabbitMQ | Product defects go to Broadcom's engineers. We prepare the evidence, work the escalation with you and run the incident until it is resolved |
| Extended support lifecycle: critical patches and CVE fixes for more release series | Which series you may run is a Broadcom decision; we plan upgrades against it |
| Enterprise security: FIPS 140-2 compliance for TLS, forward proxy support for OAuth 2.0, continuous CVE scanning of RabbitMQ and its dependencies | TLS and OAuth settings need to match the FIPS build; certificate and proxy problems show up as authentication failures |
| Disaster recovery features: continuous schema and data replication to standby clusters in external data centres, promotable after a disaster | Replication link health, standby state and promotion runbooks need rehearsing |
| Distributed shovels: shovels hosted on all cluster nodes | Shovel placement and link behaviour differ from the open-source shovel |
| AMQP 1.0 over WebSockets | Browser clients, load balancers and WebSocket timeouts become part of the support surface |
| Intra-cluster traffic compression (the page says it can reduce inter-node network load by up to 96 percent, depending on workload) | CPU use rises as network use falls; we check the trade on your nodes |
| Audit logging (on Kubernetes) | Log volume and retention; who deleted the queue questions get answers |
| Stream Browser in the management UI | Operators can inspect stream messages by offset or timestamp, so we use it in diagnosis |
Escalation Path
A cluster that stopped accepting publishers is answered before a feature-flag question. Response clocks follow plan and severity.
Clocks run by plan and severity, so blocked publishers and an upgrade-planning question are never in the same line.
P1 escalation path
A replay of one P1, from the first page to the written root-cause analysis
Response targets from 15 minutes, by plan; 24/7.
A response time, not a resolution time.
P1 opened, on-call engineer paged
Phone, email or Slack reaches the on-call senior engineer directly, whether the incident is in RabbitMQ, Kafka, Redis, PostgreSQL or Kubernetes. No web form, no tier-1 queue.
Senior engineer on your bridge
A senior engineer who already knows your environment joins a live bridge, so nobody has to re-explain your topology.
Specialists across the stack join
When an incident crosses components, the engineers who know each one join. For RabbitMQ, the path runs to the core team.
Written root-cause analysis
Root cause, the fix applied and the prevention steps, delivered after every P1.
Pricing
Tanzu RabbitMQ support typically costs $24,000–72,000 a year. The number of clusters, Kubernetes versus OVA, how many sites use warm standby and the response plan drive the quote.
Prices are in US dollars (USD). Response targets start at 15 minutes by plan, 24/7, and quote requests are answered within 1 hour. Adding other brokers or platforms to the contract can earn a multi-technology discount.
Support In Practice
Tanzu RabbitMQ support and managed-services work from AceMQ. Each card links to the full write-up: the environment, what AceMQ diagnosed or changed, and the result.
Central Bank
Moving from community support risk to enterprise-backed support with patch access, compliance posture, and procurement guidance.
IT Distribution Company
Moving from unsupported versions to supported LTS or enterprise versions with compatibility validation and rollback design.
Government IT Services Firm
Support and compliance alignment for healthcare, finance, government, and defense with supported releases, audit posture, and vendor-backed escalation.
Support options
AceMQ offers Tanzu RabbitMQ support in four tiers, each handing more of the day-to-day work to AceMQ: 24/7 support and incident management for clusters your team runs, proactive monitoring that finds quorum membership, watermark and entitlement problems first, managed Tanzu RabbitMQ in your own environment, and fully managed RabbitMQ on AceMQ's platform. Kubernetes and OVA deployments are covered.
Your platform team runs it, we take hard tickets
Best for: Teams with Tanzu RabbitMQ in production who need a senior broker engineer when the operator or the cluster misbehaves.
We check queues, certificates and upgrades ahead of time
Best for: Teams that want quorum membership, feature flags and TLS reviewed before a node replacement or a cert rotation tests them.
We operate the clusters on your platform
Best for: Organizations running Tanzu RabbitMQ on their own Kubernetes or vSphere that want specialists handling day-to-day care.
We run your Tanzu RabbitMQ clusters as a service
Best for: Application groups that want a working message broker and none of the operator, patching or paging work.
One partner for Tanzu RabbitMQ
Related material for Tanzu RabbitMQ teams: broker troubleshooting, cost against the open-source build, and Kubernetes platform support.
Broadcom provides support for the Tanzu RabbitMQ product under its subscription. Independent providers such as AceMQ support the deployment around it: Cluster Operator, quorum queues, TLS, upgrades and the platform. AceMQ works 24/7 with response targets from 15 minutes, by plan, alongside your Broadcom entitlement.
Round-the-clock response from 15 minutes on top plans, a written root cause after every P1, queue tuning, upgrade and migration work. A sample Sev 1: publishers blocked by a memory alarm, where an engineer runs rabbitmq-diagnostics memory_breakdown against the container limit. P2 is answered in 1 hour.
We work the deployment as well as the product: operators, storage, TLS and platform, with the same engineers supporting open-source RabbitMQ, Kafka and Kubernetes. Open-source RabbitMQ users can read about RabbitMQ support.
Yes. On Kubernetes we support the Cluster Operator and Messaging Topology Operator across TKG, OpenShift, EKS, AKS and GKE. On vSphere OVA we support cluster behaviour and host-level contention diagnosis, and migration from OVA to Kubernetes.
Yes. Quorum queue membership is not moved automatically when a node is replaced. Use rabbitmq-queues quorum_status to check and rabbitmq-queues grow to add a member on the new node. We check membership on every queue after node changes.
We can help you work through it with Broadcom and plan the upgrade path in the meantime. Entitlement is Broadcom's decision. We cannot grant it.
Yes, if you run a 4.x series: mirrored classic queues are not available there. We inventory queues, test consumers against quorum queue semantics, and migrate in groups.
According to RabbitMQ's commercial features page (read 17 November 2026): core engineer support 24x7, an extended support lifecycle, FIPS 140-2 for TLS, forward proxy support for OAuth 2.0, continuous CVE scanning, disaster recovery replication to standby clusters, distributed shovels, AMQP 1.0 over WebSockets, intra-cluster compression, audit logging on Kubernetes and a stream browser.
It is the commercial distribution of RabbitMQ, with additional features and support from Broadcom. The core broker is RabbitMQ. We support both under one contract.
Yes. Many teams keep a product subscription and add AceMQ for operations: quorum queues, operators, TLS and upgrade runs. You can start with a single engagement such as a queue migration or upgrade plan, then move to an annual agreement with 24/7 cover.
Contracts generally run $24,000 to $72,000 a year (USD). We scope by cluster count, deployment model (Kubernetes with the Cluster Operator or vSphere OVA), warm-standby sites, and the response plan. Subscription fees paid elsewhere are separate. Send the details and we quote within 1 hour.
Let's talk
Tell us your Tanzu RabbitMQ series and where it runs. We reply to quote requests within 1 hour.
Production incident right now?Email a Tanzu RabbitMQ Architect