Gathering Confluence Statistics to Implement DevOps Administration

Doug November 15, 2018

We're trying to set our Confluence installation and administration up in more of a devops capacity. Part of the process is to define KPIs that let us know how well the app is performing. It seems a lot of this data just needs to be pulled directly from the database via SQLs. A lot of the SQLs provided when searching seem to be related to older versions of Confluence. Even the official Atlassian docs at https://confluence.atlassian.com/disc/administration-tips/obtaining-confluence-instance-metrics are woefully out of date. An example is the SQLs referencing the ATTACHMENTS table, but that was removed in 5.7 (I think). Can anyone provide useful SQLs for modern Confluence versions (6.x perferably)? Other methods of gathering statistics are welcome, but I'd like to avoid anything that requires purchasing a plugin for that specific use case. Thanks

1 answer

Suggest an answer

Log in or Sign up to answer
0 votes
Daniel Eads
Atlassian Team
Atlassian Team members are employees working across the company in a wide variety of roles.
November 19, 2018

Hey Doug,

Let's break this out into a few different things -

Application Server (Java / DB) Performance:
An application performance monitor like NewRelic or AppDynamics will get you a ton of info about the JVM itself - heap usage, garbage collection stats, etc. in addition to basic info about the host (free/used memory, CPU, etc). It's worth checking one of these out if you're serious about performance monitoring.

Both these tools can check on what SQL queries are being run to give you hints on if there's any plugins running that might be doing things in an unoptimized way. In a previous role as an admin, I was actually able to use the DB performance tools in NewRelic to track down a single misconfigured Confluence macro that was causing an entire Confluence instance to slow down when a particular page loaded.

JMX Monitoring:
Confluence (and Jira) come with some stats enabled for export via JMX connector. You enable this in your Confluence config, and then you can attach to the Confluence server stats by running JConsole on your local machine. Check out the document for details on the types of things you can measure with JMX.

Confluence statistics:
Many of these can be pulled via the free SpaceAuditor plugin. Things you can pull include:

  • Pages in a space
  • Attachments in a space
  • Size of attachments
  • Number of contributors
  • Age of pages
  • Most modified/least modified

etc. - you get the idea!

If there are any specific Confluence-usage stats you're interested in retrieving via SQL, let me know what those are and we can work together on making sure those are 6.x-friendly queries.

Cheers,
Daniel

Doug November 21, 2018

Hi Daniel,

We're trying to gather global statistics specific to Confluence itself. I'm familiar with ways to get OS/app/db performance info, but we're being pushed to go even further and provide metrics business stakeholders can relate to. It's also to help us create a history of data to see if we can find any trends or identify issues during irregularities to be notified of issues before users encounter them. A couple of items that come to mind are:

  • Total number of newly created pages in X timeframe
  • Total number of pages updated within X timeframe
  • Total number of pages viewed within X timeframe
  • Total number of new attachments uploaded within X timeframe

I'm open to any other suggestions as to useful data relevant to the Confluence app specifically. Thanks

TAGS
AUG Leaders

Atlassian Community Events