Amazon OpenSearch Service alternatives: managed OpenSearch anywhere
OpenSearch sized for your index and run by engineers in our cloud, your cloud account or your data center, not only on AWS.
45 minutes with an engineer: we review your workload and sketch the design.
- ✓Operated by AceMQ
- ✓OpenSearch
Amazon OpenSearch Service is AWS's managed OpenSearch, billed by instance hours, storage and data transfer for domains, or by OpenSearch Compute Units for Serverless. OSStend runs dedicated OpenSearch, with node roles, shards, plugins and versions designed for your index, in AWS, another cloud or your data center. The AWS service stays the better fit for search that lives entirely inside AWS and fits its defaults.
- PostgreSQL
- MySQL
- MongoDB
- Redis
- Valkey
- OpenSearch
- Kafka
- RabbitMQ
- AWS
- Google Cloud
- Azure
- Kubernetes
- OpenShift
- Prometheus
- Grafana
- Datadog
- Dynatrace
- Splunk
Amazon OpenSearch Service and OSStend at a glance
Amazon OpenSearch Service is AWS's managed search and analytics service, available as managed clusters (domains) and as Serverless. OSStend suits teams that want the cluster designed around one index estate, or that need it outside AWS.
| OSStend | Amazon OpenSearch Service | |
|---|---|---|
| What you get | A dedicated OpenSearch cluster, designed for your index and run 24/7 | Managed clusters billed by instance hours, storage and data transfer, or Serverless billed by OCUs (AWS pricing) |
| Where it runs | Our cloud, your cloud account, your data center, Kubernetes or OpenShift | AWS |
| Tailoring | Shards, node roles, plugins and version chosen with you | Within the service's quotas and supported plugins (AWS docs) |
| Pricing | Price drivers on pricing; sized per design | Set by AWS; domain, storage and UltraWarm lines are in the pricing section below (AWS pricing) |
Sources: Amazon OpenSearch Service pricing and quotas, checked 17 Nov 2026.
Why do teams look for an alternative to Amazon OpenSearch Service?
The common reasons are quotas, plugin rules and place. AWS documents each limit, and a search estate that does not match them has to be reshaped to fit, or has to leave AWS.
| Area and managed service | Documented limit | What OSStend does instead | Source |
|---|---|---|---|
| Shard countAmazon OpenSearch Service | On OpenSearch 2.17 and above the limit is 1,000 shards per 16 GB of heap, up to a maximum of 4,000, and the default cannot be changed. On OpenSearch 1.x to 2.15 it is 1,000 and can be changed through max_shards_per_node. | What OSStend does insteadA shard plan set from your heap and index sizes. | AWS docs, checked 17 Nov 2026 ↗ |
| PluginsAmazon OpenSearch Service | Custom plugins need OpenSearch 2.15 or later, with up to 25 per account per Region, 1 GB uncompressed each, and 20 plugins of any type per domain. | What OSStend does insteadThe plugins your workload needs, agreed in the design. | AWS docs, checked 17 Nov 2026 ↗ |
| PlaceAmazon OpenSearch Service | Runs on AWS only. | What OSStend does insteadOur cloud, your cloud account, your data center, Kubernetes or OpenShift. | AWS pricing, checked 17 Nov 2026 ↗ |
Teams also compare cost, because a cluster bill grows with replica copies, cluster manager nodes and storage tiers. The OpenSearch vs Elasticsearch comparison covers the engine choice.
From our on-call engineers: how we size OpenSearch shards and master nodes
Most OpenSearch cluster instability traces back to shard count and master quorum, not to query load. We size shards before the first document is indexed, because the primary shard count of an existing index cannot easily be changed.
- ✓AWS guidance is 10-30 GiB per shard for search-latency workloads and 30-50 GiB for write-heavy log analytics, so we pick the range from the workload type first.
- ✓Amazon OpenSearch Service defaults to five primary shards and one replica per primary, while open source OpenSearch defaults to one primary, so a self-run cluster needs its template set deliberately.
- ✓We keep no more than 25 shards per GiB of JVM heap on a data node, so a node with a 4 GiB heap tops out around 100 shards.
- ✓Production clusters get three dedicated master nodes, never an even number, because two cannot elect a master after a failure and four add cost without adding failure tolerance.
- ✓If the cluster loses master quorum on Amazon OpenSearch Service, both reads and writes fail, so master placement across zones matters more than data node count.
A fair snapshot of Amazon OpenSearch Service
Amazon OpenSearch Service is the simplest way to run search beside an AWS application. If its defaults and quotas fit the workload and the data stays in AWS, there is little reason to move.
- ✓It offers two models: managed clusters, and Serverless, where compute is measured in OCUs and, on NextGen collections, scales to zero after 10 minutes of inactivity (AWS pricing, checked 17 Nov 2026).
- ✓It has storage tiers for retaining large data cheaply: UltraWarm and cold storage, with automated snapshot storage free for 14 days (AWS pricing, checked 17 Nov 2026).
- ✓It does not charge extra for Multi-AZ with Standby, per AWS's own pricing example (AWS pricing, checked 17 Nov 2026).
Full comparison: OSStend and Amazon OpenSearch Service
OSStend is listed first. The AWS column states only what AWS's pages say, with the date checked.
| Factor | OSStend | Amazon OpenSearch Service |
|---|---|---|
| Good fit | Search and log estates that need their own shard plan, plugins or version, or must run beyond AWS, including regulated data in a data center | Search that lives inside AWS and fits the service's defaults and quotas |
| Poor fit | Small indices with default settings, free tiers, self-serve sign-up | Workloads past the shard or plugin quotas (AWS docs), or that must run outside AWS |
| Where it runs | Our cloud, your cloud account, your data center, Kubernetes or OpenShift | AWS |
| Tailoring | Node layout, shards, hybrid keyword and vector search, plugins and version pinning, set in the design | Instance types, storage tiers and supported plugins (AWS docs) |
| Migration | Snapshot and restore or reindexing, as the design decides; OSStend owns the cutover | Your team plans and runs it |
| Service levels | Written SLA per plan (see service levels) | AWS publishes its own SLA; compare its current terms with the service level you need |
| Pricing model | Price drivers: size, SLA tier, where it runs. See pricing | Instance hours, storage and data transfer for clusters; OCU-hours and storage for Serverless (AWS pricing) |
Sources: Amazon OpenSearch Service pricing, quotas and custom plugins, checked 17 Nov 2026.
How does OSStend run OpenSearch?
OSStend designs the cluster around one index estate, then runs it: master and data nodes, shard plan, search features, snapshots and the version you pin. The first step is a 45-minute workload review with an engineer: the current setup is reviewed, the design is sketched, and a price follows for that design.
Managed OpenSearch is Elasticsearch as a service on the open-source fork, with keyword and vector search together. If the source is Elasticsearch rather than an AWS domain, the Elasticsearch to OpenSearch guide covers that route.
There is no fixed menu here: the plugins, 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.
Contact us if
- ✓Your index estate has outgrown the service's shard or plugin quotas.
- ✓The cluster must run in your own cloud account elsewhere, or in a data center for regulated data.
- ✓You want the shard plan, version and upgrade timing set by an engineer.
Not a fit if
- ✕Search that lives inside AWS and fits the defaults: the AWS service will cost less to start.
- ✕Spiky or intermittent search where Serverless scale-to-zero fits.
- ✕You want a free tier or a self-serve console.
What does Amazon OpenSearch Service cost?
AWS bills managed clusters by instance hours, storage and data transfer, and Serverless by OCU-hours and storage. The table reproduces the price lines in AWS's own worked example for a three-Availability-Zone production domain.
| Line in AWS's example | Rate | Usage in a month | Cost per month |
|---|---|---|---|
| 3 data nodes, r6g.xlarge.search | $0.335 per hour | 2,190 hours | $733.65 |
| 3 cluster manager nodes, c6g.large.search | $0.113 per hour | 2,190 hours | $247.47 |
| 2 UltraWarm nodes, ultrawarm1.medium.search | $0.238 per hour | 1,460 hours | $347.48 |
| EBS General Purpose SSD (gp2), 500 GB per data node | $0.135 per GB-month | 1,500 GB | $202.50 |
| UltraWarm managed storage, 1.5 TB per node | $0.024 per GB-month | 3,072 GB | $73.73 |
| Total | $1,604.83 |
US East (N. Virginia), USD, 730 hours a month. Source: pricing example 1 on Amazon OpenSearch Service pricing, checked 17 Nov 2026. Data transfer is billed separately, and AWS also lists reserved-instance and Database Savings Plans pricing. Serverless OCU rates depend on the Region and collection type; read them on the AWS page.
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.
When comparing, include replica copies, cluster manager nodes, the storage tier and cross-zone traffic on the AWS side, and the service level you need on both. Full drivers are on pricing.
What other options are there?
Several routes exist besides OSStend. The right one depends on whether the problem is control, cost, licence or place.
- Stay on Amazon OpenSearch Service.
- If the estate lives in AWS and fits the quotas, it is the simplest route. Serverless suits intermittent workloads.
- Elastic Cloud,
- sold by Elastic, for teams that want Elasticsearch itself. See the Elastic Cloud comparison.
- Run it yourself
- on your own instances, if you have engineers who will own shards, upgrades and on-call.
How do I choose?
Choose by quotas and place. If the index estate fits the service and stays in AWS, keep it; if the estate needs its own design or another place, choose dedicated OpenSearch.
- Count your shards and plugins against the AWS quotas for your OpenSearch version.
- Price the same cluster shape on both sides, including cluster manager nodes, replicas and storage tiers.
- If you need a design the service cannot give, or a place beyond AWS, choose OSStend.
How does moving off Amazon OpenSearch Service work?
The new cluster is built beside the AWS domain and loaded through snapshot and restore or reindexing, as the design step decides. Applications are tested against it before traffic switches, with the domain kept for rollback.
Review
An engineer reviews the current setup: sizes, settings, modules or plugins, versions, traffic and where it must run.
Design
The cluster design is written and signed off before anything is built: size, topology, configs, versions and upgrade calendar.
Move
The new cluster runs beside the old one and stays current while the old service stays live. Traffic switches only when the new side is verified.
Run
Named engineers operate it 24/7: backups, patching, scaling and incident response, with upgrades on the customer's change calendar.
Questions about Amazon OpenSearch Service alternatives
These are the questions teams ask before leaving, or staying on, Amazon OpenSearch Service.
What does Amazon OpenSearch Service cost?
AWS bills managed clusters by instance hours, storage and data transfer, and Serverless by OCU-hours and storage. In AWS's own example for a production domain in US East (N. Virginia) with three data nodes, three cluster manager nodes and two UltraWarm nodes, the total is $1,604.83 a month (AWS pricing, checked 17 Nov 2026).
Is there an alternative to Amazon OpenSearch Service outside AWS?
Yes. OSStend runs dedicated OpenSearch in its own cloud, in your cloud account or in your data center, with the node layout, shard plan, plugins and version designed for your index. See managed OpenSearch.
What are the shard limits on Amazon OpenSearch Service?
On OpenSearch 2.17 and above the limit is 1,000 shards per 16 GB of heap, up to 4,000, and the default cannot be changed; on 1.x to 2.15 it is 1,000 and can be changed (AWS docs, checked 17 Nov 2026).
Can I move from Amazon OpenSearch Service without downtime?
OSStend builds the new cluster beside the AWS domain and keeps both live while data is loaded and applications are tested. Traffic switches only when the new side is verified, and the domain stays for rollback.
When is Amazon OpenSearch Service the better choice?
When search lives inside AWS and fits the service's defaults and quotas, or when the workload is intermittent and Serverless scale-to-zero fits. The AWS service then costs less to start and run.
Related pages
- PricingDatabase as a service pricing: dedicated, tailoredDatabase as a service pricing from OSStend: dedicated, tailored clusters priced by engine, size, SLA tier and location. Enterprise only, with no free tier.
- ComparisonElasticsearch vs OpenSearch in 2026: which to runElasticsearch vs OpenSearch in 2026: licence, vector search, migration and when each is the better choice. Sourced and dated, by a team that runs OpenSearch.
- Service levelsService levels: two tiers, a written SLAOSStend service levels: fully managed and co-managed tiers, what each one covers, who is on call and how incidents are handled, with a written SLA for each.
- OpenSearchElasticsearch as a service on managed OpenSearchElasticsearch as a service on OpenSearch: dedicated master and data nodes, hybrid keyword and vector search and snapshots, sized for your index and run 24/7.
- Migration guideElasticsearch to OpenSearch migration: 3 methodsElasticsearch to OpenSearch migration: remote reindex, snapshot restore or Migration Assistant, the version limits that decide which, and how cutover works.
- AlternativeElastic Cloud pricing and alternatives in 2026Elastic Cloud pricing, the licence history and the alternative: managed OpenSearch under Apache 2.0, tailored to your workload and run where you choose.
Review your search estate with an engineer
45 minutes: the current domain is reviewed and the OpenSearch design is sketched.
- ✓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