A Jira instance never collapses all at once. It degrades slowly: searches slow down, boards take ten seconds to load, creating an issue becomes painful. When you diagnose it, the cause is almost always the same — an accumulation of configuration nobody governed.
In Jira, every custom field has a cost. During a search or when rendering a board, Jira must evaluate the fields present in the context. Beyond a few hundred fields, performance drops — especially fields visible across all projects ("global context").
The typical scenario: every team asks for "just one small field" for its need. Three years later, the instance has 800 fields, dozens of them duplicates — "Client", "client", "Client name", "Customer". JQL search gets slow and screens become unreadable.
The golden rule: a custom field should have a context restricted to the projects that actually use it. An unjustified global field is debt paid by the entire instance.
The discipline is: reuse before creating, restrict context, and regularly audit orphan fields (never filled, never searched).
Schemes (workflow, screens, issue types, permissions) are Jira's configuration skeleton. The structuring question is: share them across projects, or give each project its own?
The best practice for a growing organization: a small number of shared standard schemes (say a software workflow, a service workflow, a simple workflow), with documented exceptions rather than a bespoke scheme per team.
On Jira Cloud, two project types coexist, and the choice has lasting consequences.
The classic trap: letting team-managed projects multiply for their configuration convenience, then discovering you can't produce consistent organization-wide reporting because each project has its own incompatible statuses and fields. Migrating to company-managed is then costly.
A workflow should reflect the real work, not an org chart. Three principles guide healthy design.
Governance doesn't need to be bureaucratic. It needs to exist. A minimal, effective model fits in four points:
At Nimbax we often start with an instance audit: an inventory of fields, schemes and projects, spotting duplicates and unjustified global contexts. It's rarely spectacular, but it's what makes a Jira instance fast, readable and able to grow without collapsing.
