Jira is Atlassian's work tracking platform, and the single most consequential thing to understand about it is that there are two very different administration models inside one product. A company-managed project draws its behaviour from shared configuration objects — an issue type scheme, a workflow scheme, a screen scheme, a field configuration scheme, a permission scheme and a notification scheme — that may be reused across many projects. A team-managed project keeps its configuration to itself, which makes it fast to set up and impossible to standardise. Most organisational pain comes from mixing the two without deciding which is the default.
Underneath the schemes, a field is not simply present or absent. Custom fields have contexts that bind them to specific projects and issue types, screens decide where a field appears during create, edit and view, and field configurations decide whether it is required or hidden. A field that seems missing is usually present but off-screen, or out of context, and knowing which of those it is separates an administrator from a guesser. Custom field count is also the main driver of instance performance, which is why 'just add a field' is a governance decision.
The workflow engine carries the real process: statuses and transitions, conditions that decide who may move an issue, validators that decide whether the data is sufficient, and post functions that act afterwards. Automation rules have largely replaced scripting as the extension point, particularly on Cloud. And over all of it sits JQL — the query language behind filters, boards, dashboards and reports — which is the single highest-leverage skill for anyone who needs answers out of Jira rather than into it.
Why this skill matters now
Jira has quietly become a system of record. Delivery status, audit evidence, change approvals, incident history and release scope all live in it, which means its configuration is no longer an administrative detail — it is process, and in regulated organisations it is evidence. That raised the bar for whoever holds administrator rights, because a careless change to a shared scheme now has consequences in a compliance report rather than only in someone's board.
The platform also changed underneath its users. The move from self-managed servers to Cloud altered the app ecosystem, the API surface, the scripting story and the administration model, and automation rules became the supported way to extend behaviour that used to be handled with plugins or scripts. Instances that were configured to older patterns need rework, not just a lift and shift, and the rework requires understanding what the original configuration was trying to achieve.
Meanwhile the reporting demand keeps growing. Distributed programmes need status that is derived rather than assembled by hand, and every dashboard, board and report is a JQL query with a permission model behind it. Organisations are discovering that a handful of people fluent in JQL, automation and scheme design remove an enormous amount of manual reporting work — which is why those skills now appear in job specifications that never used to mention them.