01-how-to-use.md

title	expires_at	tags
How to use	never	healthchecker-release

• How to use
• Implementing in Your BOSH Release
  • 1. Vendor the healthchecker package
  • 2. Edit Monit File
  • 3. Add BPM Config
  • 4. Add Healthchecker config file
  • 5. Update job spec

How to use

healthchecker is added to a boshrelease as a monit process under the Job that is to be monitored. It is configured to perform a healthcheck against the main process in the Job. If healthchecker detects a failure, it will panic and exit. monit should be configured to run the restart-monit-job script on the failure of the healthchecker process. This script restarts the main monit process, up to ten failures in a row. After 10 consecutive failures, it gives up, since restarting the process is either in a horrible state, or the healthchecker is misconfigured and should not be causing process downtime.

Implementing in Your BOSH Release

1. Vendor the healthchecker package

Vendor the healthchecker BOSH package in your bosh release using a PREFIX. This will name the healthchecker package {PREFIX}-healthchecker in your release to prevent collisions.

git clone https://github.com/cloudfoundry/healthchecker-release.git
cd ~/PATH/YOUR_RELEASE
bosh vendor-package healthchecker ~/workspace/healthchecker-release --prefix ${PREFIX}

The PREFIX can be anything but should probably be the name of your release or a specific job in the release that will use healthchecker, e.g. cf-networking to make cf-networking-healthchecker. Once the healthchecker package has been vendored into your release, any job in your release can make use of it by following the rest of the instructions.

2. Edit Monit File

Add {JOB}-healthchecker process to the desired job's monit file, based on the example-healthchecker process in our example monit file.

1. Copy the example-healthchecker process to your monit file.
2. Change all occurances of example to {JOB}, e.g. example-healthchecker to {JOB}-healthchecker and example-healthchecker.pid to {JOB}-healthchecker.pid
3. The path /var/vcap/packages/healthchecker/bin/restart-monit-job MUST match the healthchecker PREFIX. Update the path appropriately with the PREFIX, e.g. /var/vcap/packages/{PREFIX}-healthchecker/bin/restart-monit-job.

3. Add BPM Config

Add BPM config for the new healthchecker process, based on our example bpm.yml.

1. Copy the example-healthchecker process to your BPM file.
2. Change all occurances of example to {JOB}, e.g. example-healthchecker to {JOB}-healthchecker and /var/vcap/jobs/example/config/healthchecker.yml to /var/vcap/jobs/{JOB}/config/healthchecker.yml
3. The path /var/vcap/packages/healthchecker/bin/healthchecker MUST match the healthchecker PREFIX. Update the path appropriately with the PREFIX, e.g. /var/vcap/packages/{PREFIX}-healthchecker/bin/healthchecker.

4. Add Healthchecker config file

Add a healthchecker config template to your job based on our example healthchecker config template.

1. Copy the config file to {YOUR_RELEASE}/jobs/{JOB}/templates/healthchecker.yml.erb.
2. Update the component_name to match your job, e.g. {JOB}-healthchecker.

5. Update job spec

Update the job's spec to add the healthchecker config template, healthchecker package, and any new properties.

Based on the example, you need to add:

• healthchecker config template
• package itself
  • Important: this MUST match the PREFIX used when vendoring, i.e. {PREFIX}-healthchecker
• failure_counter_file property