collective.prometheus 1.6.1

Prometheus integration for Zope/Plone.

This package publishes Plone statistics in a format that can be consumed by Prometheus.

It was largely based on munin.zope. See https://pypi.org/project/munin.zope/

It provides the following data:

• The number of running Zope ZServer threads (with the zserver extra)

• The number of Zope ZServer threads not in use (with the zserver extra)

• The number of objects in the Zope database

• Memory used by the Zope cache

• The number of objects that can be stored in the Zope cache

• ZODB load count

• ZODB store count

• ZODB connections

• Active Zope Objects

• Total Zope Objects

Installation (using Buildout)

Add collective.prometheus to your instance eggs in buildout.cfg.

Usage

Assuming Plone listens on localhost:8000, start your Plone instance and visit http://localhost:8000/@@metrics to see the output and confirm that data is being reported.

If so, add a job to your scrape_configs in pometheus.yaml:

- job_name: 'plone'
  metrics_path: '/@@metrics'
  static_configs:
  - targets: ['localhost:8000']

1.6.1

• Fix metric naming for thread-local metrics [Rudd-O]

1.6.0

• rename zope_cache_objects to zope_total_objects to better reflect what it shows

1.5.0

• Change zodb_load_count and zodb_store_count from Counter to Guage

1.4.0

• Use getActivityAnalysis() for ZODB activity

1.3.1

• make ZServer monitoring optional

• Skip ZODB activity if unavailable

1.3.0

• Zope 4 compatibility, make ZServer monitoring optional with the zserver extra

1.2.1

• Fix a typo in output

1.2

• Add documentation

• Number Zope connections instead of using their ID

1.1

• Add per connection caches active and total object counts

1.0

• Initial release