Issues

An issue is any discrete failure event — an incident, a defect, a failed check — that arrives from a connected source or is created directly. Causeloop ingests issues, runs AI analysis on each one, groups related issues into patterns, and tracks them through to resolution.

​

The Issues list

Navigate to Issues in the left sidebar to open the issues list. The list is sorted by risk score (highest first) by default.

​

Filtering and sorting

The top bar provides quick-filter tabs: For more precise filtering, use the full filter panel. You can filter by severity (p1–p4), status (one or more), source, assigned team, risk score range, pattern linkage, and date range. Valid sort fields are risk_score, ingested_at, source_created_at, severity, status, and updated_at (prefix with - for descending). Filters can be saved as Saved Views for quick recall. Select Save view from the filter panel to name and share a view with your team.

​

Bulk actions

Select multiple issues using the row checkboxes, then choose a bulk action:

• Change status — transition all selected issues to a new status
• Assign — set the assignee or team on all selected issues
• Link pattern — attach all selected issues to a pattern
• Export — download selected issues as CSV

Bulk operations are processed via POST /v1/issues/bulk.

​

Top-bar actions

• Export CSV — export the current filtered view
• Ingest event — manually create a new issue

​

Issue detail

Click any row in the list to open the issue detail page. The URL is /issues/{id}.

​

Header

The header shows:

• Breadcrumb: Issues / {external_id} (e.g. JIRA-1042)
• Status pill — the current lifecycle status (see below)
• Severity chip — P1–P4
• Source chip — the connected system the issue came from (e.g. PagerDuty, Jira, Linear)
• Timestamps — ingested at and analyzed at (UTC)

Three action buttons appear in the top bar: Assign, Link pattern, and Re-analyze.

​

Agent analysis (left column)

The left column contains the AI verdict for this issue:

​

Causal chain (left column)

Below the analysis card, the Causal chain SVG traces the contributing factors (left dots) through convergence threads to the single root-cause node (right ring). If the issue is resolved, the ring turns green and displays a checkmark.

​

Context (right column)

The right rail surfaces key metadata:

​

Similar issues (right column)

The Similar issues section lists up to three issues with high pattern-match similarity. Each row shows the external ID, a truncated title, and the similarity score (e.g. .89). Click any row to open that issue’s detail page. The similar issues data comes from GET /v1/issues/{id}/similar.

​

Issue lifecycle

Issues move through a defined set of statuses. The table below shows which transitions are allowed:

​

Triggering re-analysis

Click Re-analyze in the top bar to run a fresh AI analysis pass on an issue. You can optionally choose a model tier: fast, balanced, or deep. The issue moves to analyzing immediately and the result is written back when complete.

​

Linking to a pattern

Click Link pattern to associate this issue with an existing pattern. The issue status advances to pattern_found automatically. To remove the link, use POST /v1/issues/{id}/unlink-pattern.

​

iPad triage view

On tablet-sized screens (641–1024 px), the Issues screen shows a split layout: a scrollable queue on the left and a triage pane on the right. The queue shows only P1 in-loop issues sorted by priority. Selecting a row loads the triage pane, which displays:

• Severity and status pills
• Pattern association chip (if linked)
• Three KPI tiles: Risk score, Recurrence / 7d, Loop · 90d
• Agent analysis card with the confidence ring, cause summary, and contributing-factor chips

Action buttons in the triage pane sub-appbar are Assign and Re-analyze.

​

API reference

The Issues API is documented at /api-reference/overview. Key endpoints:

​

Related pages