Platform Documentation
Breadcrumbs

Issues and Triage

Issues are findings raised by automated background tasks that signal something in your environment needs attention. The Issues view provides a centralized place to review, investigate, and manage these findings.

Accessing the Issues View

The Issues view is the default landing screen when you open a workspace. It's accessible from the left sidebar by clicking the chat bubble icon. The view shows "Issues (N)" at the top, where N is the total count of open issues across all SLXs.

Issues List

The Issues list displays all open findings detected by background tasks, sorted by recency.

What Each Row Shows

Each row in the list represents an SLX with issues:

Element

Description

Status icon

Blue circle indicates active monitoring. Different shading indicates severity.

SLX Name

The Service-Level eXperience that detected the issue (e.g., "Checkoutservice Deployment Health")

Resource path

Full path: kubernetes/sandbox-cluster-1/online-boutique-dev/checkoutservice

Issue count

How many individual issues are grouped under this SLX (e.g., "2 issues")

Last seen

Time since the most recent occurrence (e.g., "Last 12m ago")

Expand arrow

Click to see individual issues within the SLX

  • Search bar ("Find...") — Filter issues by name or resource path

  • Status dropdown ("Open Issues") — Toggle between open, resolved, or all issues

  • Refresh button — Manually refresh the issues list

Example: Issues in the Sandbox

The Sandbox workspace typically shows issues like:

  • "Online-Boutique-Prod Namespace Health" — 2 issues

  • "Online-Boutique-Dev GitOps Flux Kustomization Health" — 1 issue

  • "Checkoutservice Deployment Health" — across dev, test, and prod

  • "Acme-Fitness Namespace Health" — 6 issues

  • "Frontend Deployment Health" — 1 issue

Issue Detail

Click the expand arrow (>) on any SLX row to see individual issues within it.

Expanded View

Each expanded issue shows:

  • Issue title — human-readable description (e.g., "Segmentation Fault in Checkoutservice")

  • Affected resource — highlighted in red when critical (e.g., Checkoutservice)

  • Occurrence count — how many times detected (e.g., "1 occurrence")

  • Last seen — when most recently detected

  • IGNORE button — red outlined button to suppress this issue

Issue Properties

Issues created by CodeBundle tasks include:

Property

Description

Severity

1 = Critical, 2 = Major, 3 = Minor, 4 = Informational

Title

Human-readable description displayed in the UI

Details

Evidence and command output explaining why the issue was raised

Reproduce Hint

The command or action that generated the finding

Next Steps

Suggested actions to investigate or resolve

Triaging Issues

Option 1: Investigate in Workspace Chat

The fastest way to investigate an issue is through Workspace Chat:

  1. Note the issue title or resource name from the Issues list

  2. Click the chat bubble icon to switch to Workspace Chat

  3. Ask about it: "What's wrong with checkoutservice in dev?" or "Tell me about the segmentation fault"

  4. The Assistant automatically finds the related issue, runs additional diagnostics, and presents a full analysis with remediation steps

Option 2: Ignore Known Issues

For expected or acceptable issues (e.g., during a maintenance window, or a known condition), click the IGNORE button on the expanded issue. Ignored issues:

  • Won't appear in the open issues list

  • Won't be surfaced by Assistants during investigations

  • Can be un-ignored later if needed

Option 3: Escalate

Issues include an escalation workflow:

  1. Click the issue to view details

  2. Use the escalation action to notify the service owner

  3. The escalation includes all issue details and diagnostic context

Issue Lifecycle

State

Description

Detected

A background task identifies a problem and raises an issue

Open

The issue appears in the Issues list with status icon

Investigated

A user or Assistant investigates via Workspace Chat

Resolved

The underlying problem is fixed; the issue clears on the next task run

Ignored

The issue is manually suppressed

Issues automatically resolve when the underlying condition clears. You don't need to manually close them — the next scheduled task run will verify the condition and remove the issue if it's resolved.

How Issues Integrate with Workspace Chat

When you ask the Assistant about a namespace or service, it automatically checks for existing issues:

  • "What's unhealthy in online-boutique-dev?" → The Assistant cites open issues for that namespace

  • It reports: "citing: 20 issues and 7 resources"

  • The response references specific issues and their severity

  • You can drill into specific issues with follow-up questions

This means you don't need to manually browse the Issues list for every investigation — the Assistant integrates issue data into its analysis automatically.

Common Issue Patterns

Pattern

Typical SLX

What to Look For

Pod crashes

Deployment Health

CrashLoopBackOff, OOMKilled, segfaults

Configuration errors

Deployment Health

Misconfigured environment variables, wrong service addresses

GitOps drift

Flux Kustomization Health

Failed reconciliation, source mismatch

Resource pressure

Namespace Health

Resource quota exhaustion, memory limits

Connection failures

Deployment Stacktrace Health

Database connection errors, service timeouts

Screenshots

Issues list — showing open issues grouped by SLX

doc-workspace-chat-main.png

Issue detail — expanded view showing individual issues and IGNORE button

64-issue-detail.png