Layer 3 · self-hosting reality check
What it actually takes to self-host OpenSearch
The docs say 2 GB. In practice you want 8 GB heap on a 16 GB machine. Here is the honest version — real requirements, real monthly cost, what you will be maintaining, and the one thing that catches people out.
Usually reached from Splunk alternatives, where OpenSearch is one of the picks.
| RAM — documented minimum | 2 GB |
|---|---|
| RAM — what it really needs | 8 GB heap on a 16 GB machine |
| CPU | 4 vCPU |
| Disk | Index is typically 1.1× the raw JSON, more with replicas |
| Monthly cost | $60–120/mo for a small cluster, against Algolia pricing that bills per search operation |
| Setup time | Half a day for one node, a week for a production cluster |
| How you install it | container or package — the Apache-2.0 fork of Elasticsearch 7.10 |
| Ongoing maintenance | Real. Shard sizing, JVM heap and index lifecycle are all yours. |
| Where it stops scaling | Petabyte scale with enough nodes. This is genuinely big-infrastructure software and it asks to be treated as such. |
The thing that catches people out
Never give the JVM more than half the machine's RAM, and never more than about 31 GB. Above 31 GB the JVM loses compressed object pointers, so a 64 GB heap genuinely performs worse than a 31 GB one. People size the heap to the box, buy a bigger box, and are baffled that it got slower.
When not to self-host OpenSearch
You want search that works without operating a cluster. Typesense or Meilisearch deliver most of the value with a fraction of the operational surface.
Every guide here carries this section. A site that only ever tells you to self-host is selling something — the useful answer is sometimes no.
Other Layer 3 self-hosting guides
Common questions
- How much RAM does OpenSearch actually need?
- 8 GB heap on a 16 GB machine in practice. The documented minimum is 2 GB, which is the figure at which the process starts rather than the figure at which it works under real use. 4 vCPU alongside it.
- What does self-hosting OpenSearch cost per month?
- $60–120/mo for a small cluster, against Algolia pricing that bills per search operation This is commodity VPS pricing and excludes your time, which is the larger cost for most people — budget for real. Shard sizing, JVM heap and index lifecycle are all yours.
- How long does it take to set up OpenSearch?
- Half a day for one node, a week for a production cluster, via container or package — the Apache-2.0 fork of Elasticsearch 7.10.
- When should I NOT self-host OpenSearch?
- You want search that works without operating a cluster. Typesense or Meilisearch deliver most of the value with a fraction of the operational surface.
- What is the most common mistake when self-hosting OpenSearch?
- Never give the JVM more than half the machine's RAM, and never more than about 31 GB. Above 31 GB the JVM loses compressed object pointers, so a 64 GB heap genuinely performs worse than a 31 GB one. People size the heap to the box, buy a bigger box, and are baffled that it got slower.