When your business critical applications slow down and have performance issues, your team complains that they see irrelevant search results, and their JQL queries return results they don’t expect, and they get bogged down in a maze of custom fields, it might mean your site is in need of a cleanup.
A frequent solution to performance issues is to invest in stronger hardware or add more nodes to your cluster. Unless you're running heavy operations on your instance or experience heavy load, getting more powerful hardware might not be the best option because it fixes the symptom and not the cause of the problem. A regular review of what’s going on on your site, and a cleanup is usually a better idea and should be a starting point of any performance assessment.
You may find that after a thorough cleanup, no additional hardware investment is necessary. Additionally, your team might feel that Jira, Confluece, or Bitbucket is more usable, they can find what they want quickly, and they can be more productive.
|Before clean up||After clean up|
Instance cleanup process
This is when you review your instance and decide on the areas to clean up.
In this stage you'll be deciding on the quick wins, communication with your team and establishing baselines.
This is when you identify the entities to clean up and run the cleanup activities.
This is when you update procedures and processes. At this point, you might also know what you can automate so that cleanup activities are less time-consuming.
In this stage you'll review the process and spend time updating it with your team. You'll also need to prepare a data maintenance plan.
|Read how to complete this stage...||Read how to complete this stage...|
This differs between applications.
It’s the amount of data in the database and the size of the Lucene index. That is why, it’s paramount to keep the index nice and small so that indexing activities take minutes not days. There might be a few causes of your Jira taking more and more time to index. Old projects, closed issues, a gazillion of custom fields, unused workflows - all these affect Jira and its indexing time. Keep them under control and you see the improvement.
But there is more. If you have heaps of unused projects, forgotten issues, and unreviewed user base, all this informations tends to pop up in search results, pickers, and browsers. This makes your team less productive and frustrated about having to waste time digging for data they should have at hand’s reach.
In Confluence there are two main types of clutter, obsolete content, such as spaces for projects and teams that have long since been wound up, and historical data, such as previous versions of pages, blog posts, and attached files, that build up over time, as the content is repeatedly edited.
The first type of content makes finding the right content difficult for your team, with search results full of old and outdated content. The second type of clutter contributes to the size of your database, index, and in the case of attached files, disk space. All of which can have an impact on Confluence performance.
Using the same instance of Bitbucket for a long time with growing and aging data can start to cause an increase of clutter that can make it more challenging for your users to find the relevant information.
What instances need cleanup
Each instance needs a bit of dusting from time to time, however the bigger the instance is and the quicker it grows, the more frequent and thorough cleanup it needs. While you can solve small instance performance issues by investing in more powerful hardware (although this proves to be a short-term solution), it is more difficult and more costly in case of a bigger instance. That’s why regular cleanups should be treated as a part of regular instance maintenance.
What’s also important is that the original cleanup should be followed by building a solid data maintenance policy. With such policy in place, every following cleanup will be less time consuming and easier. For tips on how to create such process, see the Observations and data management process section below.
Is every cleanup time-consuming
Usually the first cleanup requires some time and effort investment. However, if it becomes a regular task that’s part of maintenance and is accompanied by a strong data maintenance policy it gets quicker and no longer needs that much effort. That is why we recommend thinking about cleanup in a global way as a task that requires a thorough review of the instance but also of data. management processes and policies.
Cleanup big wins
Cleanup is not only about performance but about the teams enhancing their productivity and regaining trust in the tool they use. Quite often clutter and bad configuration results in applications not living up to their potential and users getting frustrated because they cannot achieve their goals or spend time trying to find what they want in a maze of data. If you’ve ever had to tell a team member that, yes, Jira can do what they want but not in the configuration you currently have, it’s a good indication that you might benefit from cleanup. Reducing clutter, improving performance, and enhancing teams' productivity are the three big wins of every cleanup project.