dauchande / loki-demo

Demo of Grafana Loki

loki-demo

This is a demo of Grafana Loki, showcasing how to search logs and export metrics over them.

Getting started

Clone the repo:

git clone [e-mail]:noris-network/loki-demo.git
cd loki-demo

Run locally

Prerequisites:

• Docker
• Go 1.13+

With this setup Loki, Promtail and the log_gen script will run locally, whereas Prometheus and Grafana will run inside Docker containers on the host network.

Run make install to build ts_gen and download Loki and Promtail:

make install

Next start the services:

Each of the below commands should be run in a separate terminal window.

make run/loki
make run/promtail
make run/docker/up
make run/log_gen

If you have an AWS S3 bucket, you can store loki's chunks in it with.

make run/loki/s3 ACCESSKEY=<your aws access key>  SECRETKEY=<aws secret key> \
     S3ENDPOINT=<s3 endoint>  BUCKETNAME=<bucket name>

Using it

Access Grafana via http://localhost:3000 and add the following datasources:

Datasource	Name	URL
Prometheus	Prometheus	http://localhost:9090
Loki	Loki	http://localhost:3100
Prometheus	Loki as Prometheus	http://localhost:3100/loki

As of Grafana 6.4 the Loki as Prometheus datasource is necessary to use aggregation functions like rate or count over LogQL results.

You can then import the sample dashboard in dashboards/log_gen.json:

Going on Explore on the left hand side lets you evaluate the logs via LogQL:

{job="demo_log", service="api", level="error"} |= "cpu"

As of Grafana 6.4, LogQL functions need to be sent agaist the Loki as Prometheus datasource.

sum by (handler) (rate({job="demo_log", handler!=""})[5m])

Cleanup

First stop all services that run in the foreground:

CTRL+c

Then shutdown the docker-compose stack:

make run/docker/down

Running make clean will remove the log file created by log_gen, all binaries, created Docker volumes and all data created by Loki and Promtail:

make clean