What is LynxDB?
LynxDB is an open-source log analytics database written in Go. The same binary queries local files and standard input or runs as a standalone server. Both modes use OpenSearch Piped Processing Language (PPL).
One binary, two ways to run it
| To do this | Use |
|---|---|
Inspect a log file or aggregate the output of kubectl logs | Pipe mode |
| Ingest logs continuously and query them over HTTP | Server mode |
Files and pipes
lynxdb query --file access.log 'source=main status>=500 | top 10 uri'
The CLI loads the input into a temporary index named main and removes it
when the command exits. It keeps JSON fields and extracts timestamps,
key=value pairs, and fields from formats such as combined access logs
and syslog.
Server
Start lynxdb server in one terminal. In another terminal, ingest a file and
query the stored events.
lynxdb ingest access.log
lynxdb query 'source=main status>=500 | top 10 uri'
Each index is a columnar store partitioned by time. Token filters let the
server skip blocks that cannot match a text search. Besides the CLI, the
server accepts logs through NDJSON, Elasticsearch _bulk, OTLP, Splunk HEC,
and syslog.
The query language
source=main level=ERROR earliest=-1h
| stats count() as errors, p95(response_time) as p95 by host
| sort - errors
A query names an index, then pipes events through commands such as
where, stats, and sort. LynxDB compiles PPL to an intermediate
representation, optimizes the plan, and runs it on the columnar engine.
See supported features for command and function
coverage, or PPL errors for diagnostics.
Next steps
- Quick start to run a query on sample data.
- CLI reference for commands and output formats.
- REST API for HTTP clients.
- Architecture for storage and execution.