LogJuicer Roles

LogJuicer extracts anomalies from log files.

You have two options to integrate LogJuicer in your Zuul job post-run phase:

Deploy the web-service and use the report-logjuicer.
Use the run-logjuicer to create the report locally:

For single-node job, the role can be used in untrusted playbooks.

For multi-node job, the role must be used on localhost (the executor) between the fetch-output and the log upload. Note that a future version may support merging reports produced on multi nodes through an untrusted playbooks.

In both case, a new artifact named “LogJuicer Report” will be provided to access the build report.

report-logjuicer

Create a LogJuicer report link

This role emits an artifact to create a report through a LogJuicer web service. Add the following task to your job post-run phase:

- when: not zuul_success | bool
  include_role:
    name: report-logjuicer
  vars:
    logjuicer_web_url: https://softwarefactory-project.io/logjuicer
    zuul_web_url: https://zuul.opendev.org/t/{{ zuul.tenant }}

Role Variables

logjuicer_web_url: The http url of the LogJuicer web service.

zuul_web_url: The http url of zuul-web. For multi-tenant deployment, add /t/{{ zuul.tenant }}.

run-logjuicer

Create a LogJuicer report

This role runs the LogJuicer tool to create a report of the current build. For single-node jobs, the role can be used on the test instance in the job post-run phase.

- when: not zuul_success | bool
  include_role:
    name: run-logjuicer
  vars:
    zuul_web_url: https://zuul.opendev.org/

For multi-node jobs, the role must be used on localhost (the executor) between the fetch-output and the log upload.

Role Variables

zuul_web_url: The zuul-web URL, to lookup baselines.

logjuicer_max_run_time Default: 900: Maximum runtime in seconds.