Regular maintainer tasks#
This document summarises and collects links to information about tasks other than opening new code contributions that maintainers should consider, especially if they already have several PRs with reviews requested.
Review the issue backlog and perform some of the regular backlog maintenance tasks
Pick up the oldest, unassigned implementation plan
This will eventually result in a new PR but usually takes a while to complete and are easy to overlook in favour of direct code contributions
If an older implementation plan is assigned and hasn’t been started and you could pick it up, consider pinging the current assignee to offer to take if off their plate
Search through Sentry for older issues that might be getting worse; see section below for details
Read and respond to active discussions
Research unknown/under-documented parts of the project; see section below for advice
Pick up active incident investigation
Check with MSR if you need help identifying what incidents need help investigating
Proofread documentation; see the documentation review page for advice
Sentry review#
While the Openverse Mini-Support Rotation[1] covers recording and triaging new Sentry issues, it does not include regularly checking the frequency of existing issues. You can review Sentry issues by frequency by visiting the following links:
If you see an issue with heaps of events and see that the GitHub issue is not being worked on or has been prioritised such that it will not be picked up soon, use your best judgement to determine whether it needs to be worked on immediately (it could be the next code contribution you work on) or if it’s safe enough to just add to the weekly dev chat agenda.
Unknown/under-documented parts of the project#
There are a handful of issues in our backlog that request documentation for parts of the project. You can find those by scanning the list of text-aspect issues. To identify other parts of the project that might need additional documentation, think of the following:
Is there anything that you know particularly well that you feel other maintainers might not?
On the other hand, is there anything that you do not know well and feel could be better documented? This could be a good time to learn more about it with hands on experience researching and documenting it. This is especially important if it is something that none of the maintainers know well.
Is there ambient context or domain-specific knowledge that is not explicitly documented that should be?
If you start writing documentation, you’re likely to open a PR. This is okay, text/documentation PRs are excluded from the list of requested reviews as they have a significantly lower review burden than code changes, in particular because the consequences of incorrect or poor documentation are usually far less severe than incorrect code.