Cassandra

Managed Apache Cassandra in your cloud or data center

Apache Cassandra with the topology, replication and compaction set for your workload: dedicated clusters in OSStend's cloud, inside your own cloud account, or co-managed in your data center, with engineers on call 24/7.

45 minutes with an engineer: we review your workload and sketch the design.

  • Operated by AceMQ: 160+ customers, 10+ years
  • 24/7 on-call engineers

Managed Cassandra from OSStend is for teams that want Apache Cassandra itself, not a compatible service, and want someone to own repairs, compaction, node replacement, backups, multi-datacenter layout and upgrades to 5.0. Engineers design the cluster, move you in and run it 24/7.

160+AceMQ customers
10+Years running these systems
24/7Engineers on call
3Places it runs: our cloud, yours, your data center
Engines we run, on the platforms you already use
  • PostgreSQL
  • MySQL
  • MongoDB
  • Redis
  • Valkey
  • OpenSearch
  • Kafka
  • RabbitMQ
  • AWS
  • Google Cloud
  • Azure
  • Kubernetes
  • OpenShift
  • Prometheus
  • Grafana
  • Datadog
  • Dynatrace
  • Splunk
Built for your workload

What OSStend tailors for Cassandra

A Cassandra cluster is designed from the data model it will carry: partition keys, read and write mix, consistency levels, retention and the datacenters it spans. The design fixes node count, replication, compaction and where it runs, and you sign it off before anything is built.

AreaSet for your workload
Size and shapeNode count and size, disks and the token layout per node (the documentation calls num_tokens 256 a reasonable default), chosen from data size and traffic (topology docs, checked 18 Nov 2026).
Replication and multi-datacenterReplication factor per keyspace and per datacenter, and the consistency levels clients use, so a datacenter can be lost without losing writes the design promised to keep.
CompactionThe compaction strategy per table, matched to the access pattern, and the disk headroom compaction needs.
Versions and upgradesThe Cassandra release your drivers are validated on. The project lists 5.0, 4.1 and 4.0 as maintained lines (download page, checked 18 Nov 2026); upgrades follow your change calendar.
Where it runsOSStend's cloud, your own AWS, Azure or Google Cloud account, your data center, Kubernetes or OpenShift, or an air-gapped site.
MonitoringMetrics and alerts, including dropped messages, pending compactions, repair status and disk use, forwarded to your Datadog, Dynatrace, Splunk or Prometheus.

There is no fixed menu here: the topology, consistency settings and cluster size your workload needs are written into the design you sign off before anything is built, and that design is what OSStend runs.

What we run 24/7 for Cassandra

Cassandra's day-to-day work is repair, compaction, node lifecycle and backups. Skipped repairs are what turn into deleted data coming back, so that schedule is the first thing the on-call team owns.

  • Repairs. Repair every node inside the tombstone grace period; the Cassandra docs give a default gc_grace_seconds of 10 days and advise a repair at least every 7 days, with incremental repair as the default and full repairs run occasionally (repair docs, checked 18 Nov 2026).
  • Compaction. Watching pending compactions and disk headroom, and changing strategy when a table's access pattern changes.
  • Node replacement. Replacing a dead node with the documented replace procedure, and repairing afterwards when the node was down longer than the hint window (topology docs).
  • Backups. Snapshots made with hard links, with incremental backups switched on where the design needs them (they are off by default), and the schema kept alongside because a table backup does not include DDL (backup docs, checked 18 Nov 2026).
  • Multi-datacenter. Replication, consistency levels and repair scope per datacenter, and the failover drill that proves one datacenter can carry the load.
  • Upgrades, 4.x to 5.0. A rolling, one-node-at-a-time plan with a snapshot first and a rollback path; see the section below.
  • Monitoring and incidents. Alerts forwarded to your tools, with a runbook per alert and engineers on call 24/7.

Upgrading Cassandra 4.x to 5.0

Cassandra 5.0 is the latest generally available line, with 4.1 and 4.0 still maintained. An upgrade is a rolling one, node by node, and the release notes list changes that need action before you start.

Java.
The 5.0 notes say Java 8 is removed and Java 11 is the lowest supported version; Java 17 is listed as experimental there (NEWS, checked 18 Nov 2026).
Compaction.
Tables on DateTieredCompactionStrategy must move to TimeWindowCompactionStrategy first, because the old strategy was removed.
TLS settings.
If cassandra.yaml relied on the default keystore and truststore password, set the passwords explicitly or startup fails.
Rolling and schema.
Upgrade one node at a time so the cluster converges on one schema version, and take a snapshot before you start, because snapshots from the previous major version need sstableupgrade before sstableloader can use them.
Defaults that changed.
For example disk_access_mode now defaults to mmap_index_only, so the design review lists each changed default against your configuration.

The release notes are long; the design review checks your cassandra.yaml, tables and drivers against them and fixes the path from your current version. We make no claim about an upgrade we have not rehearsed on your cluster.

From our on-call engineers: the Cassandra checks in our runbook

These come from the Apache Cassandra documentation, read on 18 Nov 2026, and are the checklist we build a new cluster's runbook from. They are documentation, not a record of clusters we have run, and we quote no customer figure.

  • Repair is scheduled, not remembered: a node that goes more than the grace period without one risks deleted data reappearing.
  • Incremental repair does not protect against disk corruption, operator error or bugs, so the schedule includes occasional full repairs.
  • A replacement node that takes longer than the hint window to bootstrap is followed by a repair.
  • After adding nodes, nodetool cleanup runs on the nodes that lost ranges, because Cassandra does not remove that data by itself.
  • A snapshot does not carry the schema, and incremental backups are off by default, so restore is rehearsed from the artefacts the design actually produces.
  • Decommission is for a live node leaving; removenode is for a dead one. Mixing them up is an avoidable mistake, so the runbook names each.

Cassandra services compared: Astra DB, Amazon Keyspaces, Instaclustr and others

Other ways to get Cassandra range from compatible serverless services to managed Apache Cassandra on other platforms. They differ in whether it is Cassandra itself, where it can run and who is on call. Each fits a different buyer.

Illustrative: plan names and shapes stand in for any provider's menu.
OptionWhat it isWhere it fits
DataStax Astra DBA cloud database described by its vendor as built on Apache Cassandra, now sold under IBM (product page, checked 18 Nov 2026).You want a vendor-run Cassandra-based database with vector search and no cluster to operate
Amazon KeyspacesA serverless, Apache Cassandra-compatible service from AWS; AWS says you can run existing Cassandra application code against it (documentation, checked 18 Nov 2026).You are on AWS, want no servers to manage and your code fits its supported features. It is compatible, not Apache Cassandra itself
InstaclustrManaged Apache Cassandra on AWS, Azure and Google Cloud, in the cloud or on premises, with an availability SLA it publishes (product page, checked 18 Nov 2026).A Cassandra-first estate, or a buyer who wants a NetApp contract. See Instaclustr alternatives
AivenAiven's page states that its Cassandra service has been discontinued (Aiven, checked 18 Nov 2026).Not an option for new clusters
OSStendApache Cassandra itself, designed per workload and run 24/7 in OSStend's cloud, your cloud account or your data centerYou want upstream Cassandra, in your own account or data center, with engineers on call

OSStend is our service and is listed once. Each row uses only what the vendor publishes, and we state no price. Amazon Keyspaces or a vendor's serverless plan is the better choice when its feature set covers your workload and you want nothing to operate.

How it works

Design. Move. Run. for Cassandra

For Cassandra, an engineer designs the topology and compaction and you sign it off, OSStend moves the data across with both clusters live, and named engineers then run it 24/7.

  1. Design

    An engineer reviews your data model, read and write mix, consistency levels, datacenters and drivers, and writes the cluster design: nodes, replication, compaction, backups and where it runs. You sign it off before anything is built.

  2. Move

    OSStend builds the new cluster and moves data across, by adding it as a datacenter and rebuilding, or by loading snapshots, with both sides live until the new one is verified. Applications switch in stages.

  3. Run

    Named engineers run it 24/7: repairs, compaction, node replacement, backups, upgrades and incident response, with the commitment written into your contract.

Deployment

Deployment and service levels

Cassandra runs dedicated in OSStend's cloud, inside your own cloud account, or co-managed in your data center. The first two are fully managed; the third is co-managed, with OSStend engineers operating Cassandra on infrastructure you own.

AceMQ engineersOne on-call team, one contract, wherever it runs
  • Dedicated in OSStend's cloud

    Single-tenant Cassandra clusters on Google Cloud or AWS, with nodes spread across zones, or datacenters across regions.

    Fully managed
  • In your cloud account

    Cassandra inside your cloud account and VPC, so applications keep private paths and the infrastructure bill stays with you. AWS first, then Azure and Google Cloud.

    Fully managed
  • In your data center

    Nodes on your hosts, Kubernetes or OpenShift, air-gapped sites included, run through an operator with supported charts and images.

    Co-managed

Fully managed deployments carry an uptime commitment. Co-managed deployments carry a 24/7 response-time commitment, because you own the infrastructure. The service levels page sets out the tiers, and deployment shows each option in detail.

Uptime and response figures are set per plan and written into your contract. AceMQ, which operates OSStend, sets its support response targets from 15 minutes, by plan.

Operated by AceMQ

Proof from AceMQ's operating record

OSStend is operated by AceMQ, a company with 160+ customers, 10+ years of running open-source infrastructure and engineers on call 24/7. Those company facts match acemq.com.

Operated by AceMQCompany figures, as published on acemq.com
160+AceMQ customers
10+years operating
24/7on-call engineers

AceMQ names customers only with their written permission, and this page makes no claim about past Cassandra deployments. OSStend's Cassandra service is a designed offering backed by AceMQ's operating practice for other engines.

Pricing

Pricing for managed Cassandra

Pricing for managed Cassandra follows four drivers: the engine, the cluster size, the SLA tier and where it runs. An engineer prices the specific design after a workload review. There is no free tier and no card sign-up.

Engine
each engine has its own operating profile, and several engines share one contract and one on-call team.
Cluster size
node count and size, disk and retention, which follow data volume and traffic.
SLA tier
fully managed or co-managed, and the uptime or response-time commitment attached.
Where it runs
OSStend's cloud, your cloud account or your data center.

Prices are quoted per design, after the workload review, so the figure is for the cluster you need rather than a plan size. Send the engine, rough size and where it must run, and the engineer prices that design.

The pricing page shows the model that applies across every engine.

Enterprise only

When to contact OSStend, and when not to

Contact OSStend when your Cassandra estate has outgrown a standard plan or has to run somewhere a plan cannot. If the workload fits a standard plan, or you want a free tier or card sign-up, a turnkey service is the better choice.

Contact us if

  • You run Apache Cassandra and want engineers on call for repairs, compaction and upgrades.
  • Cassandra has to run in your own cloud account or your data center.
  • You are leaving a Cassandra-compatible or vendor service and want upstream Cassandra.
  • You want Cassandra, Kafka and PostgreSQL under one contract and one on-call team.
Talk to an engineer

Not a fit if

  • A serverless Cassandra-compatible service covers your workload and you want nothing to operate.
  • You want a free tier, or to sign up with a card. For Cassandra, OSStend begins with a conversation with an engineer.
  • You will keep running Cassandra yourself and only want advice: speak to AceMQ.
FAQ

Questions about managed Cassandra

These are the questions teams ask before handing Apache Cassandra to a managed service.

Talk to an engineer →

What is managed Cassandra?

Apache Cassandra run for you: cluster design, repairs, compaction, node replacement, backups, upgrades and monitoring, with engineers on call 24/7. OSStend designs the cluster for your workload first, so it is not a fixed plan.

Is OSStend affiliated with the Apache Cassandra project?

No. OSStend is not affiliated with or endorsed by the Apache Software Foundation. Apache Cassandra is open source under the Apache License 2.0, and OSStend runs it for customers.

Can Cassandra run in my own cloud account or data center?

Yes. Nodes run inside your cloud account and VPC, or on your hosts, Kubernetes or OpenShift in a data center, co-managed by OSStend engineers.

How often does Cassandra need repair?

Cassandra's documentation advises repairing every node at least once every 7 days, because the default tombstone grace period is 10 days. Incremental repair is the default and full repairs run occasionally. The schedule is part of the design.

How do you upgrade from Cassandra 4.x to 5.0?

Node by node, after a snapshot and a review of the release notes: Java 11 is the lowest version in 5.0, DateTieredCompactionStrategy tables move first, and TLS passwords are set explicitly. See upgrading Cassandra 4.x to 5.0.

How is this different from Amazon Keyspaces or Astra DB?

Those are vendor-run services, one Apache Cassandra-compatible and one built on Cassandra. OSStend runs Apache Cassandra itself, where you choose: our cloud, your cloud account or your data center.

What does managed Cassandra cost?

The price follows node count and size, disk, the SLA tier and where it runs. The pricing page sets out the model, and an engineer prices the design after a workload review.

Trademarks and affiliation

Apache Cassandra and Apache Kafka are trademarks of the Apache Software Foundation. OSStend is not affiliated with or endorsed by the Apache Software Foundation, and the project names are used only to say which software we run. Amazon Keyspaces is a trademark of Amazon, and DataStax Astra DB of its owner.

Talk to an engineer

Talk to an engineer about your Cassandra estate

45 minutes with an engineer: bring your data size, consistency settings and where Cassandra runs now, and the engineer sketches the cluster design.

  • A design sketch for your workload, not a plan menu
  • A zero-downtime cutover path for what you run today
  • A price for that design, after the review
See pricing →

Your request goes to the OSStend team. No newsletter sign-up.

Get a quoteTalk to an engineer
2Engines
3Where should it run?
4What do you need?

Your request goes to the OSStend team. No newsletter sign-up.