Skip to main content

@Jenkinsci analytics with @Elasticsearch and #Kibana

I have been doing a lot of work with presenting Jenkins data in dashboards. Jenkins by default does not come with a built-in database. All the data gets stored in the filesystem. However, Jenkins has a REST API that allows us to read data as json. I searched around the Googlesphere to see if anyone had figured out how to capitalize on this REST API for performing analytics and found almost nothing directly related to this.

I was searching for open to public Jenkins instances to see how other people use Jenkins. I came across Elasticsearch's Jenkins site and I was pleasantly surprised. Elasticsearch is a json document storage application. There is a Jenkins plug-in that directly inputs data to Elasticsearch. 

I implemented this proof of concept to perform Jenkins analytics with Elasticsearch as the document store and Kibana as the analytics engine. I automated the installation part using Docker. There are better ways to automate this. The steps given below do not require any knowledge of Docker. 

Pre-requisites:

Vagrant
Oracle VirtualBox
Docker client installed on a running Ubuntu Vagrant box

Step 1: Create the docker containers

Execute the following docker commands to create the Docker containers needed for each of the apps.

#Create Elasticsearch container
docker run --name elasticsearch --hostname elasticsearch -p 9200:9200 -p 9300:9300 -d elasticsearch

#Create Jenkins container
docker run -p 8080:8080 -p 50000:50000 -v $HOME/jenkins_home:/var/jenkins_home --link elasticsearch:es --name jenkins -d jenkins

#Create Kibana container
docker run -e ELASTICSEARCH_URL=http://elasticsearch:9200 -p 5601:5601 --link elasticsearch:es --name kibana -d kibana


Step 2: Port forwarding from VirtualBox to host

Expose the following ports between the host computer and the virtual machine. Optionally, you may choose to add this to the Vagrantfile so that the port forwarding is automatically set whenever the Vagrantbox is started.

8080:8080

9200:9200

5601:5601


Step 3: Install the Jenkins Logstash plug-in

Click on Manage Jenkins and Manage Plug-ins



Click on 'Available' and Filter for 'Logstash'. Checkbox and click install without restart.

Step 4: Configure Jenkins to push data to Elasticsearch

Open http://localhost:8080/configure from a browser on the host computer. Look for the 'Logstash Plugin' section and fill out the following details.

Indexer type: ELASTICSEARCH
Host name: http://localhost
Port: 9200
Username: 
Password: 
Key: /logstash/type

Step 5: Create a Jenkins job and configure

Create a new jenkins job.


Check 'Send the console log to Logstash'.


Add a build step 'Execute shell'.


Step 6: Run Jenkins job

Run the job and see the console log.





Step 7: Verify data is getting to Elasticsearch

Open http://localhost:9200/logstash/_search?pretty in a browser and search for the jobname in the json.

Step 8: Get started with Kibana

I am still trying to figure out Kibana. Check out the visualize section and try to create a graph. I'll post that on another blogpost.

Open 'http://localhost:5601' in a browser and try it out.



Labels:

Popular posts from this blog

Create #VirtualPrivateCloud, NAT Instance and NAT Gateways on @AWSCloud

Create a Virtual Private Cloud, NAT instance and the new NAT Gatweay ... and making it all work. This is a YouTube playlist of three videos.
Read more

Cheat sheet to create a #VPC and Subnets on @AWSCloud

One of the critical things to remember for working with a AWS VPC is creating and using it. I had hard time remembering how to do it, so, I wrote down a cheat sheet for myself.  If anyone wants to follow along, just navigate to the VPC page on the AWS Console and start with 'Create VPC' button. Please note that this may cost some dollars if you are not on the free tier. If you are on the free tier and make mistakes, it may cost some dollars. In the steps below, we will be creating the following on a new VPC: An internet gateway One public subnet with routes for accessibility from the internet One private subnet without any routes One EC2 web server with Apache installed in it and serving a sample html page - using the public subnet. One EC2 server with the private subnet and security group that allows access to resources running on the public subnet only.  Create VPC Name tag: myVPC CIDR Block: 10.0.0.0/16 Tenancy: default (Must have default. Otherwise, i...
Read more

SQL Server tables with JSON

We can indeed store json data as-is into a traditional Microsoft SQL Server database. The document hosted on Microsoft's site left a lot of questions and unknowns that I had to explore and experiment to figure out the right recipe for creating a table to store json, inserting the data as json and querying for the values of individual keys within the json. Here you go:  --Create a table with an identity column and a nvarchat(max) column to store the individual json documents create table dbo.logs (     _id bigint primary key identity,     json_log nvarchar(max) ); --Add a constraint to the json_log column of the table to ensure that the table accepts only json as value to store ALTER TABLE dbo.logs ADD CONSTRAINT [json_log record should be formatted as JSON] CHECK (ISJSON(json_log)=1); --Insert json into the table insert into dbo.logs values ('{"key": "value"}'); insert into dbo.logs values ('{"key": "value1"}'); --Query for al...
Read more