Clover is a root cause analysis system. There are three separate systems contained in this same repo. These would normally belong in separate repos, but are contained here for simplicity.
Clover analyzes a lot of data, And it relies on the unpredictabily provided only by a running system. For this reason, I've built a simple, but practical example system to run and produce data.
This example system consists of a load balancer, 2 web servers, a MySQL server, and an Elasticsearch cluster consisting of 5 nodes (3 master, 2 data)
Metric extractors are set up on this system using Telegraf, sending the data to InfluxDB. You can also set up Grafana to view these metrics.
Now that we have an example system providing us with data, we can use Clover to transform, evaluate, alert, and give insights via a report on this data.
Clover is meant to run via Apache Spark. You can find the instructions for installing Spark in the Spark docs
After cloning this repo, run sbt clean assembly
This will create a jar file in the target directory. This jar will be used for actually running Clover.
Clover requires configuration in order to know which metrics to run. One key aspect of Clover is that you don't have to understand the metrics, but you do have to tell Clover where to find them. There are examples in clover/src/main/resources/.
In order to run the transformer, run:
$SPARK_HOME/bin/spark-submit \
--class "clover.service.TransformMetrics" \
path/to/clover/jarfile.jar all
where all is the name of the configuration file
$SPARK_HOME/bin/spark-submit \
--class "clover.service.TrainAlgorithm" \
path/to/clover/jarfile.jar all
where all is the name of the configuration file
$SPARK_HOME/bin/spark-submit \
--class "clover.service.EvaluateMetrics" \
path/to/clover/jarfile.jar all
where all is the name of the configuration file
Clover Web is a Ruby on rails application for managing alerts and reports.
To run Clover Web, cd into the directory clover-web and run:
bundle install
rake db:migrate
rails server
Before reporting will work, we need to download the most recent jar file from https://s3.amazonaws.com/master-project-clover/clover.jar and place it in clover-web/bin/clover.jar
Alerting is run via rake task. In the terminal, cd to clover-web and run:
rake run_alerting_service