1. Overview #
An integrated stack.
The building blocks #
ClickStack is the observability stack offered by ClickHouse: a preconfigured OpenTelemetry Collector, ClickHouse with an optimized schema, and HyperDX, a search and analysis interface acquired by ClickHouse. It aims for an experience close to commercial tools, in open source.
2. Installing #
All-in-one or separate.
All-in-one #
docker run -d --name clickstack \
-p 8080:8080 -p 4317:4317 -p 4318:4318 \
docker.hyperdx.io/hyperdx/hyperdx-all-in-one
# Interface : http://localhost:8080 (création du compte au premier accès)The ingestion key (API key) shown in the team settings must be sent by the SDKs in the authorization header.
export OTEL_EXPORTER_OTLP_ENDPOINT=http://localhost:4318
export OTEL_EXPORTER_OTLP_HEADERS="authorization=<cle_ingestion>"
export OTEL_SERVICE_NAME=frontProduction #
In production, the building blocks are separated: replicated ClickHouse (or ClickHouse Cloud), Collectors sized for ingestion, and HyperDX connected to ClickHouse. A Helm chart is provided.
Common mistake: The all-in-one image stores everything in a single container: perfect for a trial, unsuitable for data you want to keep.
3. Using HyperDX #
Searching and correlating.
Features #
| Feature | Usage |
|---|---|
| Search | Lucene-like syntax (level:error service:checkout "timeout") or SQL |
| Traces | Timeline, spans, logs associated with each span |
| Sessions | Browser session replay (with the browser SDK), linked to back-end traces |
| Charts and dashboards | Aggregations on any field |
| Alerts | On searches and charts |
Good practice: Going from a log to its trace, then to the browser session that triggered it, is the stack's strong point: instrument front and back end with OpenTelemetry to benefit from it.
