minitasks.txt

Jenkins versions: https://hub.docker.com/r/jenkinsci/jenkins/tags/
Graphite container: https://github.com/hopsoft/docker-graphite-statsd

# AWS minitasks
- Fix the SSH key dissemination to gitserver (at least for AWS, volume mounts suck)
- Jenkins: similar way to disseminate SSH key for gitserver, also add a volume mount for keys
- Jenkins mount local volume, i.e. EFS
    + EFS standup basis
- (BIG): Kubernetes-based standup 
X Set up gitserver to use port 2222 for Jenkins (local and remote)
    - Problem: when running linked in containers, gets port 22 not 2222
    - Maybe build args?
X Use this script to use instance store for the workers: https://gist.github.com/kamermans/176c190670a163b1147e

# Minitasks
X Jenkins server, with plugins.txt and jenkins-home
X Jenkins server with custom plugins
X Jenkins server with local jenkins_home
X Git server
X External build agent(s)  - https://github.com/jenkinsci/docker-ssh-slave

# Fixes
X Fix Jenkins user setup for container (use my special scripts)
    + Alternately, every time do a chown -R jenkins:jenkins $JENKINS_HOME
X FIX the git checkout from repo: https://github.com/jkarlosb/git-server-docker
    + git credentialsId: 'jenkins-new',url:'ssh://git@gitserver:22/git-server/repos/sample.git'  (probably invalid repo structure)
X Agents need to be auto-created by the swarm plugin
X Agents need to have their SSH keys and hostnames prepopulated on Jenkins
X Fix the JENKINS_HOME getting wrong user permissions (sigh, again, probably cause of update ctr issues)
- Generate GC log for Jenkins in docker, written to /tmp, for some reason the normal suggested settings will not accomplish this, when added to `ENV JENKINS_JAVA_OPTIONS` (may need other variable)
- Why does linux disk stats report a bunch of 'nb*' devices, and why don't the aggregate metrics look correct in docker -- write Junit testcase with Assume to restrict to running in linux environment

# Polish
X Automate agent setup with Docker swarm
NOPE Docker compose what I can using docker swarm?
X Look at reducing the graphite metrics update interval and submitting filtered metrics to get better #s

# Basics
X random-job-builder -- seems to take a while to kick in
X connect agents
X metrics
X metrics reporting -- working on Graphite integration: http://graphite.readthedocs.io/en/latest/config-carbon.html
   --> just point the jenkins instance at graphite:2003 with prefix jenkins and we're done!, UI at localhost:81
X Clean up metrics reporting to give fixed or total numbers, not histograms

# Features
X Linux metrics plugin for IO and network:
    + Add vmstats - wio, nice, system, cpu steal, etc
    + Add iostat reporting
    + Use https://www.kernel.org/doc/Documentation/ABI/testing/procfs-diskstats for raw figures on IO
    + Network stats: https://www.kernel.org/doc/Documentation/ABI/testing/sysfs-class-net-statistics
    + Also: /sys/block/vda/stat and /sys/block/vda/vda1/stat
X Random job builder: filter for jobs
X Random job builder: keep queue full
X Select metrics of interest to report on?

# Metrics key:
RX2JW0L9ZvRI2LH6enXPBldWa9841H2oaiUS38cEDNgmJIpsPEy06BRqCgyrDuxR

URL: http://localhost:8080/metrics/RX2JW0L9ZvRI2LH6enXPBldWa9841H2oaiUS38cEDNgmJIpsPEy06BRqCgyrDuxR/metrics?pretty=true

Need to create metrics key, and then hit API

# InfluxDB Data Retention
X Set it up to store all the things

## Image
https://github.com/hopsoft/docker-graphite-statsd


## Graphite customization, need to set aggregation of 1s or more, 90d retention
in /opt/graphite/conf/storage-schemas.conf