Layer 3 · self-hosting reality check
What it actually takes to self-host HyperDX
The docs say 4 GB with 2 cores — the README's stated minimum for testing. In practice you want 8 GB. Four services share the box — ClickHouse, the HyperDX app, the OpenTelemetry collector and MongoDB — and ClickHouse's own sizing guide says total memory should not go below 8 GB.. 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 Sentry alternatives, where HyperDX is one of the picks.
Wondering whether you need to at all? Is Sentry free? — what the free tier actually allows, and where the wall is.
| RAM — documented minimum | 4 GB with 2 cores — the README's stated minimum for testing |
|---|---|
| RAM — what it really needs | 8 GB. Four services share the box — ClickHouse, the HyperDX app, the OpenTelemetry collector and MongoDB — and ClickHouse's own sizing guide says total memory should not go below 8 GB. |
| CPU | 2 cores to test. ClickStack's production guide sizes 1 vCPU per 10 MB/s of sustained raw ingest, plus more for query concurrency. |
| Disk | ClickStack's sizing assumes about 10x compression on raw telemetry, so 10 GB a day of raw logs lands as roughly 1 GB a day on disk. Retention sets the total — and the default is 3 days. |
| Monthly cost | $48/mo for an 8 GB, 4 vCPU VPS at DigitalOcean list price (read 2026-09-30), less on budget hosts; storage grows with retention |
| Setup time | 10 minutes with the all-in-one image; a day to secure it, set retention and wire OpenTelemetry into your apps |
| How you install it | docker compose (ClickHouse, HyperDX, OTel collector, MongoDB) for a single-server deployment; the all-in-one image is for testing and loses data unless /data/db, /var/lib/clickhouse and /var/log/clickhouse-server are mounted |
| Ongoing maintenance | ClickHouse and MongoDB upgrades, TTL tuning as volume grows, and securing ingestion with the API key from Team Settings. HyperDX is now built by ClickHouse as part of ClickStack, so follow the ClickStack docs on clickhouse.com. |
| Where it stops scaling | Docker Compose on one server is fine for production where fault tolerance is not required, in ClickHouse's own words. Past that, run ClickHouse as a cluster or move to managed ClickStack on ClickHouse Cloud; HyperDX can point at any existing ClickHouse. |
The thing that catches people out
By default, ClickStack keeps data for 3 days. The TTL applies to otel_logs, otel_traces, the metrics tables and hyperdx_sessions (session replay), so the error from last week is simply gone when you go looking for it. Set the ttl on the collector's ClickHouse exporter before the first ingest, or ALTER TABLE … MODIFY TTL on tables that already exist — and size the disk for the retention you choose.
When not to self-host HyperDX
You only want crash reports and alerts for one small app, and nobody on the team wants to run ClickHouse. GlitchTip accepts Sentry's SDKs on a much lighter stack, or stay on Sentry's hosted plan.
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 HyperDX actually need?
- 8 GB. Four services share the box — ClickHouse, the HyperDX app, the OpenTelemetry collector and MongoDB — and ClickHouse's own sizing guide says total memory should not go below 8 GB. in practice. The documented minimum is 4 GB with 2 cores — the README's stated minimum for testing, which is the figure at which the process starts rather than the figure at which it works under real use. 2 cores to test. ClickStack's production guide sizes 1 vCPU per 10 MB/s of sustained raw ingest, plus more for query concurrency. alongside it.
- What does self-hosting HyperDX cost per month?
- $48/mo for an 8 GB, 4 vCPU VPS at DigitalOcean list price (read 2026-09-30), less on budget hosts; storage grows with retention This is commodity VPS pricing and excludes your time, which is the larger cost for most people — budget for clickHouse and MongoDB upgrades, TTL tuning as volume grows, and securing ingestion with the API key from Team Settings. HyperDX is now built by ClickHouse as part of ClickStack, so follow the ClickStack docs on clickhouse.com.
- How long does it take to set up HyperDX?
- 10 minutes with the all-in-one image; a day to secure it, set retention and wire OpenTelemetry into your apps, via docker compose (ClickHouse, HyperDX, OTel collector, MongoDB) for a single-server deployment; the all-in-one image is for testing and loses data unless /data/db, /var/lib/clickhouse and /var/log/clickhouse-server are mounted.
- When should I NOT self-host HyperDX?
- You only want crash reports and alerts for one small app, and nobody on the team wants to run ClickHouse. GlitchTip accepts Sentry's SDKs on a much lighter stack, or stay on Sentry's hosted plan.
- What is the most common mistake when self-hosting HyperDX?
- By default, ClickStack keeps data for 3 days. The TTL applies to otel_logs, otel_traces, the metrics tables and hyperdx_sessions (session replay), so the error from last week is simply gone when you go looking for it. Set the ttl on the collector's ClickHouse exporter before the first ingest, or ALTER TABLE … MODIFY TTL on tables that already exist — and size the disk for the retention you choose.