Skip to main content

Vector to LynxDB

If you already run Vector​

Keep the Elasticsearch sink and point endpoints at LynxDB.

 sinks:
lynxdb:
type: elasticsearch
- endpoints: ["http://elasticsearch:9200"]
+ endpoints: ["http://lynxdb:3100"]
mode: bulk

Full annotated config​

# Vector configuration for LynxDB.
sources:
files:
type: file
include:
- /var/log/*.log

sinks:
lynxdb:
type: elasticsearch
inputs: [files]
endpoints: ["http://lynxdb:3100"]
mode: bulk
bulk:
index: "vector-logs"
compression: zstd
# LynxDB's bulk endpoint accepts zstd-compressed requests.

Where the events go​

Each event is stored in the LynxDB index named by the sink's bulk.index (here vector-logs). Vector's default index, vector-%Y.%m.%d, creates one index per day; query them together with source=vector-*.

lynxdb query 'source=vector-logs | stats count() by file'

Each event is stored as sent: the file source's file and host fields arrive as fields, timestamp sets the event time, and message holds the log line. Extract more from it at query time; see Field Extraction.

Troubleshooting​

Vector's health check sends GET / and GET /_cluster/health to the endpoint; LynxDB answers both. If the sink reports 400 responses, check that the index name uses only lowercase letters, digits, ., _ and -.

Supported behavior​

Supported: mode: bulk, index and create bulk actions, zstd and gzip compression, health checks, and _index routing to LynxDB indexes.

Not supported: Elasticsearch query APIs, and update or delete actions.