Which Companies Offer Apache Druid Support?

AceMQ offers Apache Druid support: 24/7 cover with a 15-minute emergency response, for self-managed Apache Druid on any version. The other sources are the Apache Druid community, Imply, and consultancies such as Deep.BI, compared below.

The project does not publish a support window for older releases, so a Druid cluster that has not been upgraded for a year or two is, in practice, on its own.

What the Apache Druid project supports

At the time of writing Apache Druid 38.0.0, released on 1 October 2026, is the latest stable release on the Apache Druid download page, and older releases sit in the Apache archive. The project's versioning documentation describes how major, minor and bug-fix versions relate and recommends running the stable release, but it does not promise fixes for older lines. Community support means the user mailing list, Slack, GitHub issues and Stack Overflow, with no response time and no access to your cluster.

What is Apache Druid good for, and is it a data warehouse?

Druid is an open-source, real-time analytics database built for sub-second OLAP queries over event data at high concurrency. It ingests streams from Apache Kafka or Amazon Kinesis and batches from object storage or HDFS, stores data as time-partitioned segments with optional rollup, and answers aggregation queries through Druid SQL over the HTTP API or JDBC. It powers user-facing dashboards, product and network analytics, and time series workloads where many users query fresh data at once.

It is not a general-purpose data warehouse. Druid's strength is fast, filtered aggregation on recent events; large joins between big tables, frequent updates to individual rows and ad hoc reporting across many unrelated tables fit a warehouse better. Its main disadvantage is operational: a Druid cluster is several cooperating services, Coordinator, Overlord, Broker, Router, Historical and MiddleManager or Indexer, plus deep storage, a metadata database and service discovery, and each one can be the bottleneck.

Imply and other commercial Druid options

Imply sells Imply Enterprise, a commercial distribution of Druid that you deploy, and Imply Polaris, a cloud database service built on Druid. Both come with Imply's support for Imply's software. Consultancies such as Deep.BI offer Druid tuning and support services. If you run open-source Apache Druid yourself, on virtual machines or Kubernetes, the vendor route means migrating to a vendor distribution first.

The Apache Druid incidents that actually page people

• Kafka ingestion lag. The supervisor falls behind because there are too few ingestion tasks for the partitions, task slots on the MiddleManagers are exhausted, or handoff to Historicals is slow, and dashboards show stale data. The fix depends on which of those it is, and the offset lag metrics tell you.
• Query performance that degrades over months. Too many small segments from streaming ingestion with no compaction, poor partitioning, or rollup disabled where it would help. Each query fans out across thousands of segments and the Brokers struggle under concurrency.
• Historicals that cannot load segments. Disk or memory on the Historical tier is full, load rules ask for more replicas than the tier can hold, and the Coordinator reports unavailable segments, so queries silently return partial results.
• Metadata store and deep storage problems. The metadata database grows or slows, or deep storage permissions change, and ingestion tasks fail at publish time even though the data was read correctly.
• Upgrades across several releases. Configuration properties, extensions and default behaviours change between major versions. Skipping several releases at once without reading every set of upgrade notes is how a routine upgrade becomes an outage.

Where independent Apache Druid support fits

• Self-managed Apache Druid on any version, including releases well behind the current one.
• Performance and reliability work, such as compaction, partitioning, rollup and tiering, that a break-fix contract alone does not cover.
• The streaming pipeline in front of it, usually Apache Kafka, because ingestion problems often start upstream.
• Planned upgrades through several releases, tested against your own ingestion specs and queries first.

AceMQ provides 24/7 Apache Druid support with a 15-minute emergency SLA and named senior engineers, and Kafka support for the ingestion side. Examples: an ingestion lag remediation and query performance support.