AutoMQ/automq
Diskless Kafka® on S3. 10x Cost-Effective. No Cross-AZ Traffic Cost. Autoscale in seconds. Single-digit ms latency. Multi-AZ Availability. observed · 2026-08-28
Health v2 · maintenance only
95/100
- Activity 99
- Release rhythm 98
- Longevity 79
How is this computed?
round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10) — computed 2026-09-02. Adoption (stars, forks) is never an input.
- gap_med: 9.5
- age_days: 1112
- days_rel: 15
- days_push: 7
- n_releases_24m: 55
Adoption not part of the score
10574 stars · 761 forks observed · 2026-08-28
What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded
AutoMQ is a cloud-native, 100% Apache Kafka-compatible streaming platform that replaces broker-local disks with S3-compatible object storage in a shared-storage, diskless architecture. It offers sub-10ms P99 latency, seconds-level autoscaling of stateless brokers, zero cross-AZ replication traffic, and significantly lower infrastructure costs than traditional Kafka deployments.
Use cases
- replace apache kafka with a cheaper cloud-native alternative
- run kafka on s3 object storage instead of ebs volumes
- reduce kafka cross-availability-zone data transfer costs
- autoscale kafka brokers in seconds without partition reassignment
- migrate existing kafka clients and tools with zero code changes
- deploy kafka on aws, gcp, or azure with unlimited retention
- stream events into a data lake like iceberg for analytics
- self-host a managed kafka service in my own cloud account (byoc)
When to choose
- you run Kafka on cloud infrastructure and want to cut storage, compute, and cross-AZ traffic costs
- you need elastic scaling without capacity planning, partition reassignment, or over-provisioning
- you want Kafka protocol compatibility so existing producers, consumers, Kafka Connect, Flink, and Spark keep working unchanged
- you need long or effectively unlimited retention without managing disks
- you want to deploy on Kubernetes (EKS, AKS, GKE) or via Docker with minimal ops overhead
When to avoid
- you run Kafka entirely on-premises with no access to S3-compatible object storage
- you depend on Kafka features or ecosystem tools that assume shared-nothing local-disk brokers
- you need strict single-digit-millisecond latency for every access path, since S3-backed reads can be slower than local NVMe
- you prefer a fully open-source managed-free stack without any commercial BYOC/managed service component
- your workload is small enough that plain Apache Kafka or a managed service like MSK is already cost-effective
Facets
service · maturity active
message-queue streaming cloud self-hosted deployment big-data cloud-computing microservices analytics cloud self-hosted jvm kafka diskless s3 event-streaming kafka-compatible object-storage serverless multi-az cost-optimization apache-kafka messaging data-engineering real-time docker kubernetes
8 sources
- readme: https://github.com/AutoMQ/automq · fetched 2026-08-28 · fd74e83c4fc1
- homepage: https://www.automq.com · fetched 2026-08-29 · b32221cfeb47
- site_page: https://docs.automq.com · fetched 2026-08-29 · 9e292ee59980
- site_page: https://www.automq.com/about · fetched 2026-08-29 · c63a6676c399
- site_page: https://www.automq.com/pricing · fetched 2026-08-29 · eaa23e57ce23
- site_page: https://www.automq.com/pricing?latency=high&avg=200&peak=200&partitions=2000&retention=72&az=multi&fanout=3&requests=12800&compare=mskstandard&kafka_fetch=leader&kafka_net=50&kafka_disk=50&mskexp_fetch=leader&mskexp_net=50&mskstd_fetch=leader&mskstd_net=50&mskstd_disk=50 · fetched 2026-08-29 · eaa23e57ce23
- site_page: https://www.automq.com/the-diskless-engine · fetched 2026-08-29 · a88a2565001a
- site_page: https://www.automq.com/faq · fetched 2026-08-29 · d82a0b93aab5
Member repositories
| Repository | Role | Health v2 |
|---|---|---|
| AutoMQ/automq | main | 95 |
For agents
Data as of 2026-08-30T08:39:29.467469+00:00 · Report a problem