Emergency Incident Response
Database unreachable, commits hanging, or a failover that didn't happen. With response targets from 15 minutes, by plan, a senior PostgreSQL DBA joins a live bridge with access to diagnose — not a ticket acknowledgement.
Tanzu Postgres support
Last reviewed by Tyler Eastridge
AceMQ provides 24/7 support for VMware Tanzu Postgres: the PostgreSQL server builds and the Kubernetes operator that runs them. Most calls start with one of three problems: a failover that did not trigger, transaction ID wraparound warnings, or connection storms every time pods roll. We cover the database and the platform it sits on.
AceMQ is Broadcom's 2025 Tanzu Partner of the Year, a Leading Broadcom Tanzu partner and a Leading Global Services Partner. Our senior DBAs run PostgreSQL on Kubernetes under the Tanzu operator — autovacuum falling behind on freeze, connection storms after a rolling restart, a synchronous standby blocking every commit, and checkpoint stalls that are really PVC IOPS limits. Every ticket reaches a named engineer who already knows your cluster.
Planning a build, upgrade or migration instead? See AceMQ Tanzu Postgres consulting and services.
Customers




Partners


Incident Triage
Vacuum, connection, commit and failover problems on Postgres running under the Kubernetes operator, with the log line or catalog view we read first.
| Symptom you're seeing | What we check first | Typical resolution |
|---|---|---|
| Log warns the database must be vacuumed to avoid wraparound, or it refuses commands | SELECT datname, age(datfrozenxid) FROM pg_database against autovacuum_freeze_max_age (200 million by default), then which table holds the oldest relfrozenxid and why autovacuum is not finishing it | Remove the blocker (long transaction, abandoned replication slot, prepared transaction), then run VACUUM FREEZE on the table |
| "Too many connections" whenever pods roll | pg_stat_activity grouped by application and state, against max_connections, and the pool size times replica count of each deployment | Put a pooler in front, cap pool sizes, stagger rollouts |
| Commits hang on the primary while it looks healthy | pg_stat_replication: with synchronous replication a missing or slow synchronous standby holds commits. Also check wait_event in pg_stat_activity | Repair or replace the standby; review synchronous_standby_names and the failure policy |
| Write latency spikes on Kubernetes at regular intervals | Checkpoint frequency and duration in the log (log_checkpoints), WAL volume, and storage class latency during the spike | Spread checkpoints with max_wal_size and checkpoint_completion_target; move to faster volumes |
| Primary unhealthy but the operator promoted nothing | Operator events and instance status with kubectl describe on the Postgres resource, and whether a replica was eligible (lag, health) | Repair the eligible replica, correct the resource spec, then retest failover |
| Sequential scans and disk growth on a table that should be small | Dead tuples in pg_stat_user_tables, autovacuum settings per table, long-running transactions holding back cleanup | Per-table autovacuum thresholds; remove the long transaction |
| Point-in-time recovery fails, or backups are incomplete | Restore into a scratch instance and replay to a target time; check WAL archive continuity | Fix archiving, schedule restore tests and keep the proof |
Each P1 closes with a written cause that points to the table, setting or volume at fault.
Deployment coverage
| Deployment | What AceMQ covers | What Broadcom covers | Excluded |
|---|---|---|---|
| Tanzu Postgres on Kubernetes (VMware SQL with Postgres for Kubernetes operator) | 24/7 incident response, RCA on P1s, instance and operator behaviour, failover, vacuum and query tuning, backup and PITR assurance | Broadcom ships and supports the commercial builds and operator for entitled customers | Patched Tanzu builds |
| Tanzu Postgres on Tanzu Platform | Database layer: configuration, backup, performance | Broadcom supports the platform integration | Platform faults |
| Tanzu Postgres on VMs (non-operator) | Same database scope | Broadcom supports the build | Hardware and OS faults outside scope |
| Community PostgreSQL next to Tanzu | Same engineers; see /postgresql/support/ | Community project, no vendor | Extensions without an owner |
| Connection poolers and application clients | Pool sizing, timeouts, retry behaviour | Not applicable | Application code is reviewed, not rewritten |
Better together
Broadcom publishes the Tanzu Postgres server and operator releases and supports them for entitled customers. AceMQ adds 24/7 database engineers who also understand the Kubernetes layer, so one team owns an incident that spans the database and the platform.
| Broadcom subscription | AceMQ engineers alongside it | |
|---|---|---|
| What it covers | The product, its defects and releases | The database, the operator behaviour and the workload |
| Hours and response | Per the terms of your subscription | 24/7 across Americas, EMEA and APAC; response targets from 15 minutes, by plan |
| Community PostgreSQL | Not covered | Covered |
| 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
Query and vacuum tuning, operator behavior, storage performance and backup recovery are handled together for Tanzu Postgres clusters.
Database unreachable, commits hanging, or a failover that didn't happen. With response targets from 15 minutes, by plan, a senior PostgreSQL DBA 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 vacuum policy, pool configuration, or storage class change that prevents recurrence. Delivered as standard, not on request.
EXPLAIN ANALYZE plan review, index strategy across B-tree, GIN, GiST, and BRIN, autovacuum sizing for your write pattern, partitioning for large tables, and pgbouncer pool configuration that matches your workload.
Tanzu Postgres operator upgrades, instance CR changes, storage class and PVC sizing, pod anti-affinity for zone-aware placement, and health checks that actually detect an unhealthy primary.
WAL archiving verified end to end, restore testing that performs a real restore on a schedule, retention policy review, and documented recovery runbooks with measured recovery times rather than estimates.
As Broadcom's 2025 Tanzu Partner of the Year, we handle Tanzu for PostgreSQL subscription questions alongside the technical ones: entitlement reconciliation, true-up exposure, renewal terms, and quotes inside 24 hours.
Why AceMQ
AceMQ holds Broadcom's top VMware partner award for the Americas in 2025. That means an escalation path into Tanzu engineering when a problem is genuinely an operator defect, and one team covering both your support and your subscription.
The same senior DBAs stay on your account. They know your instance topology, your storage class, and your write profile — so a P1 starts with diagnosis instead of you explaining your cluster while commits are hanging.
You reach a senior PostgreSQL DBA directly by phone, email, or Slack. No help desk collecting information to pass along, no escalation approval between you and someone who can read pg_stat_activity against your storage metrics.
Postgres on Kubernetes fails in ways bare-metal Postgres does not: PVC throughput limits, probes that misjudge health, pods rescheduled mid-checkpoint. We diagnose the database and the platform together, because splitting them across two vendors is how incidents stretch into days.
Round-the-clock coverage for customers in 26+ countries. Your 3am replication stall is someone's mid-afternoon — no overnight skeleton crew, no waiting for a region to wake up.
Quarterly health checks covering freeze age, bloat, slot retention, and backup verification — plus shared intelligence across our support base when a version-specific bug surfaces on someone else's cluster first.
Postgres on Kubernetes and OpenShift, vSphere with Tanzu, Cloud Foundry and public cloud clusters, plus sites with no outbound access.
Version coverage
| Tanzu Postgres line | Notes (Broadcom advisories, searched 17 Nov 2026) | AceMQ coverage |
|---|---|---|
| 18.x | Latest advisory found covers 18.4.0, dated 2 Jun 2026 | 24/7 support |
| 17.x | 17.10.0 in the same advisory | 24/7 support |
| 16.x | 16.14.0 in the same advisory | 24/7 support |
| 15.x and 14.x | 15.18.0 and 14.23.0 in the same advisory | 24/7 support, with an upgrade plan |
| Kubernetes operator | Documentation set 3.0 found; operator-to-server pairing not confirmed | 24/7 support; upgrade runs the operator and instances together |
In practice
A failover that does not trigger usually means no replica was eligible. We check, in order: whether the replica was healthy and within its lag limit at the time, whether the operator saw the primary as unhealthy (or a probe still passed), whether storage or networking trapped the pod in a state that is neither up nor down, and whether the resource spec asked for the number of instances you thought. The fix is rarely a restart. It is a repaired replica, a corrected probe or a spec change, followed by a failover test in a window you choose.
Escalation Path
Commits that hang on a healthy-looking primary need a DBA who thinks about synchronous standbys and PVC latency at the same time.
Clocks follow your plan, so a database refusing writes is not queued behind a tuning request.
Pricing
Tanzu Postgres support typically costs $24,000 to $72,000 a year. The total follows the number of instances and clusters, how much HA and pooling you run, and the response plan you pick.
Prices are in US dollars (USD). Response targets start at 15 minutes by plan, 24/7, with quote replies inside 1 hour. Adding other data services to the same contract can lower the price.
Support In Practice
Tanzu Postgres support work from AceMQ. Each card links to the full write-up: the cluster, what AceMQ diagnosed or changed, and the result.
Regional Payments Processor
Finding why a Tanzu Postgres primary stayed in service while refusing queries, and fixing the health probe so the operator promotes a replica.
Software-as-a-Service Analytics Vendor
Support engagement tracing afternoon write latency spikes on Tanzu Postgres to volume burst credits, not the database settings everyone had been tuning.
Public Sector Records Agency
Restoring a Tanzu Postgres backup into a clean namespace for the first time, and finding that WAL archiving had been failing quietly for weeks.
Regional Energy Distributor
Sequencing a Tanzu Postgres operator upgrade and a PostgreSQL major-version upgrade across four clusters, rehearsed on cloned data.
Support options
AceMQ offers Tanzu Postgres 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 freeze age, archive failures and weak probes first, managed Postgres on your own Kubernetes, and fully managed Postgres on AceMQ infrastructure. Licensing is quoted in the US only, with a 72-core minimum.
Your platform team runs it, our DBAs assist
Best for: Teams running Postgres on Kubernetes who want a senior DBA for vacuum, failover or commit latency incidents.
Regular reviews of vacuum, poolers and backups
Best for: Teams that want freeze age, connection limits and restore tests checked on a schedule rather than during an outage.
We operate your Postgres clusters on your platform
Best for: Organizations keeping data on their own Kubernetes or vSphere estate who want DBA operation provided for them.
AceMQ runs Postgres for you as a service
Best for: Application teams that want a dependable Postgres backend without building in-house database operations.
One partner for Tanzu Postgres
Our DBAs also support the other Tanzu data services and plain PostgreSQL outside the operator.
Broadcom supports the commercial product for entitled customers. AceMQ provides independent 24/7 support for the database and the Kubernetes layer, with response targets from 15 minutes, by plan.
Treat it as urgent: left alone the database stops accepting commands. Find the oldest relfrozenxid, remove what blocks autovacuum and freeze that table.
Yes. See /postgresql/support/.
When failover does not trigger, we check what the operator saw: health probe results, replication state and whether a synchronous standby was available to promote. Often a probe passes while commits hang. We fix the detection logic, then rehearse a controlled switch.
Yes: cap pool sizes, add a pooler and stagger rollouts.
Restore to a scratch instance and replay to a target time. We schedule and evidence it.
Cover runs at all hours, with response from 15 minutes on the fastest plans. Example Sev 1: the database refuses commands and warns of transaction ID wraparound. A DBA joins you, runs a vacuum freeze on the oldest tables and sends a written cause. P2 1 hour, P3 4 hours, P4 next business day.
Yes. We read your entitlement against the Postgres instances and clusters you run and report any mismatch. We do not sell licences.
Yes. A backup and recovery test, a vacuum tuning exercise or a version upgrade can be scoped and quoted on its own, with no yearly commitment. If incidents keep recurring, an annual agreement adds 24/7 cover and keeps the same DBAs working on your clusters.
Contracts commonly fall between $24,000 and $72,000 a year (USD). We look at instance and cluster count, HA topology, pgbouncer tiers, backup and recovery requirements, and your response plan. Send the cluster details and a quote returns within 1 hour.
Let's talk
Send the Tanzu Postgres version, instance count and cluster layout. We reply within 1 hour.
Production incident right now?Email a Tanzu Postgres Architect