A ready-to-import Kibana dashboard for logs shipped by
ElasticsearchTransport.
It only works because that transport writes Elastic Common
Schema
fields (@timestamp, log.level, service.name, …) alongside LogQuill’s
own — Kibana never ingests data itself, it only visualizes what’s already
in Elasticsearch in a shape it recognizes.
logquill-dashboard.ndjson is a Kibana saved-objects
export
containing:
logquill-*), time field @timestamp.log.level,
service.name, and message, newest first.log.level over
time (an error-rate-at-a-glance view), and a breakdown of records per
service.name (LogQuill’s logger name).Point ElasticsearchTransport at an index matching logquill-* (the
default index="logs" won’t match the bundled index pattern — either
name your index e.g. logquill-app, or edit the index pattern’s title
after importing to match whatever you actually use):
from logquill import ElasticsearchTransport, Logger
transport = ElasticsearchTransport(url="http://localhost:9200", index="logquill-app")
logger = Logger("app", transports=[transport])
logquill-dashboard.ndjson.date_histogram/terms aggregations
on a histogram/pie viz) to rebuild in a couple of minutes by hand
using the field names above.log.level.keyword/service.name.keyword assumes
Elasticsearch’s default dynamic mapping created those .keyword
sub-fields (true out of the box for most setups). If your index has a
custom mapping without them, point the visualizations at whatever your
equivalent keyword field is.logquill trace/logquill serve (see the
CLI section of the README), which read JSONL or
SQLite directly and don’t need Elasticsearch/Kibana at all.