SYSTEM Cited by 3 sources
Jira¶
Jira is Atlassian's issue-tracking and project-management product. Issues live under Projects (also called spaces in the newer product taxonomy), can be grouped into Epics, and carry labels for cross-cutting tagging. Jira's workflows define the state machine (Open → In Progress → Done, with customisable transitions). The cloud-hosted multi-tenant deployment is Jira Cloud.
Role on this wiki¶
Jira appears on the wiki at three altitudes:
- As a downstream sink for automated tickets — the canonical alert- channel-to-Jira auto-ticket workflow: an automated signal (accessibility violation, monitoring alert, security scan) is routed through a Slack workflow that creates a pre-populated Jira ticket with a canonical label and placement, collapsing the alert-to-triage step into a single button click.
- As a multi-tenant SaaS exhibiting configuration-bloat failure modes at scale — Jira Cloud as the host platform whose limits and guardrails plus optimisation tools make it the canonical wiki instance of the configuration-bloat problem and the optimisation- tools answer architecture.
- As a workflow-substrate for AI agents — the canonical wiki instance of the work-item-as-agent-prompt framing. Jira's structured custom fields encode the agent's task brief; Jira's workflow state machine carries the human approval gate; Jira's workflow transition automations carry per-procedure custom system prompts that trigger the agent. Atlassian's first-party Add an agent to workflow transitions feature is the integration surface; Atlassian's own Jira team used it for KTLO automation (stale-flag cleanup, flaky-test remediation) producing 500+ merged PRs in 70 days on the Jira repo.
Jira Cloud limits and guardrails (per space)¶
As Jira scales to "100k+ users and beyond", Atlassian introduced hard caps on configuration entities to protect core experiences (view, search). Per Atlassian's data limits and guardrails, the published caps include:
- Up to 700 custom fields per space.
- Up to 150 work types (formerly issue types) per space.
- Up to 20,000 field options per field.
- Up to 50 issue security levels per space.
- Up to 50 grants per permission.
- Plus "limits around versions, workflows, components and priorities."
(Source: sources/2026-05-14-atlassian-optimisation-tools-for-jira-reducing-configuration-bloat)
Optimisation tools architecture¶
Hard limits create a corollary problem: how do tenants already over the cap get back under without breaking their setups? Atlassian's answer is the Jira Optimisation Tools — admin experiences plus async backend workflows that:
- Report unused / low-value configuration entities per space, computed asynchronously by the Pre-computation Framework.
- Remediate via opinionated bulk actions consuming those reports (dissociate unused fields, split field configuration schemes, clean up unused work types).
- Surface all of this through the Site Optimiser admin dashboard.
The reporting + remediation contract is detailed in systems/atlassian-jira-optimisation-tools and the underlying batch-orchestration platform in systems/atlassian-precomputation-framework.
Seen in¶
- sources/2026-05-14-atlassian-optimisation-tools-for-jira-reducing-configuration-bloat (2026-05-14, Atlassian) — first-party architectural retrospective on the Jira Cloud Optimisation Tools. Names Jira's specific limits, the optimisation-tools
- Site-Optimiser surface, the Pre-computation Framework's Initialise → Scan-steps → Finalise contract, the polymorphic-usage-tables decision (rejecting per-entity-type tables to avoid "millions of tables across production"), the Memcached + DB tiered storage split, the ≈70% utilisation-prioritised refresh strategy, and the audit-log-as-rollback v1 contract. Aggregate outcome numbers: tens of millions of unused fields and work types removed; some customers streamlining hundreds of thousands of fields and work types in a single day.
- sources/2025-01-07-slack-automated-accessibility-testing-at-slack
— Slack's accessibility team maintains a Jira workflow
for triaging a11y violations. A Slack alert-channel
workflow spins up a Jira ticket from an
accessibility-automation alert with label
automated accessibility, placed into the accessibility-triage Jira Epic. Verbatim surface: "A new JIRA bug ticket, A11YAUTO-37, was created." Canonicalises the alert-as-ticket-creation-UI pattern. - sources/2026-06-01-atlassian-how-we-cut-up-to-80-of-engineering-chores-using-ai-agents-in — canonical wiki instance of Jira as agent-orchestration substrate. Atlassian's Jira team uses Jira's custom-fields, workflow state machine, and workflow-transition automations as the substrate for delegating KTLO chores to AI agents: daily heuristic crons emit one work item per stale flag (patterns/heuristic-cron-emits-agent-work-items); engineers transition status to delegate-to-agent (patterns/jira-status-transition-triggers-agent-workflow); the transition automation passes a custom system prompt; the agent loads a per-codebase or per-failure-category specialist skill (patterns/agent-skill-with-fallback-chain, patterns/test-category-classifier-then-specialist-skill) and opens a draft PR. Reported throughput: 500+ merged PRs in 70 days on the Jira repo from the stale-flag cleanup workflow alone; ~80% reduction in flaky-test eng hours from the flaky-test workflow. First wiki articulation of Jira as agent-prompt-substrate (concepts/work-item-as-agent-prompt).
Related¶
- systems/atlassian-jira-optimisation-tools — the configuration-bloat remediation tools.
- systems/atlassian-jira-site-optimiser — the admin dashboard surface.
- systems/atlassian-precomputation-framework — the async batch-reporting platform underneath.
- systems/rovo-dev — likely consuming agent surface for the workflow-transition-triggered KTLO automation.
- systems/atlassian-teamwork-graph — cross-product knowledge graph that supplements work-item context.
- concepts/configuration-bloat — the failure mode being addressed at the multi-tenant-SaaS altitude.
- concepts/work-item-as-agent-prompt — the substrate framing at the agent-orchestration altitude.
- concepts/agent-as-first-pass-investigator — the human-in-the-loop discipline for agent-authored draft PRs.
- concepts/ktlo-engineering-chores — primary work category delegated through Jira-as-substrate.
- patterns/alert-channel-to-jira-auto-ticket-workflow — the canonical pattern Jira is the sink for at the alert-pipeline altitude.
- patterns/jira-status-transition-triggers-agent-workflow — Jira-Cloud-native agent integration mechanism.
- patterns/heuristic-cron-emits-agent-work-items — upstream detection layer that emits agent-ready Jira work items.
- patterns/agent-skill-with-fallback-chain — per-codebase skill dispatch that runs inside the agent the workflow transition fires.
- patterns/test-category-classifier-then-specialist-skill — per-failure-category dispatch.
- companies/atlassian