Skip to content
Ileen Docs

Monitoring Active Jobs

Whenever Ileen runs a background operation — generating a plan, executing a coding task, deploying a service — a job is created and tracked. The Active Jobs indicator gives you a real-time view of everything running across all your projects.

The Active Jobs indicator

The indicator lives in the sidebar footer, above the user avatar button. It shows a live text label:

  • “N jobs running” — spinning loader icon; at least one job is actively executing
  • “N jobs pending” — jobs are queued, waiting for an agent to become free
  • “No active jobs” — everything is idle (dot icon)

Click the label to expand the dropdown. The dropdown groups all jobs by project and updates the elapsed-time counter every second while it is open. Click anywhere outside to close it.

Job types by mode

🎨 Design Mode

Label in the UIAgentTriggered by
Processing ConversationArchitectSending a message to the AI Architect
Generating PlanArchitectClicking “Generate Plan”
Generating MilestonesTech LeadTech Leader producing the initial milestone set
Regenerating MilestonesTech LeadClicking “Regenerate milestones”
Iterating MilestonesTech LeadChatting with the Tech Leader to refine milestones
Generating TasksTeam LeadTeam Leader breaking milestones into tasks
Starting DevelopmentCoding AgentClicking “Save & Start Development”
Developing Task #xxxxxxxxCoding AgentEach individual coding task running

💻 Develop Mode

Label in the UIAgentTriggered by
Executing CodeCoding AgentA Task Queue task or Task Plan execution

📝 Content Mode

Content operations (briefing, social posts, changelog) run inline and typically complete within the chat turn — they appear in the Active Jobs dropdown only for longer background operations.

🚀 Deploy Mode

Label in the UIAgentTriggered by
Analyzing Codebase for DeploymentDeploy AgentStarting the deploy pipeline
Analyzing CodebaseDeploy AgentAgent re-analysing a specific step
Deploy AssistantDeploy AgentEach conversation turn with the Deploy Agent
Validating GateDeploy AgentAgent validating a deployment gate result
Analyzing DeploymentDevOpsDevOps agent analysing the infrastructure plan
Refining DeploymentDevOpsIterating on the deployment configuration
Iterating DeploymentDevOpsChatting with the DevOps agent mid-deployment

Reading the job dropdown

Each job entry shows:

  • Status icon — clock (pending/queued), spinning loader (running), check (completed), warning triangle (failed)
  • Label — human-readable description of what the agent is doing (see tables above)
  • Agent name — which agent is responsible: Architect, Tech Lead, Team Lead, Coding Agent, Deploy Agent, or DevOps
  • Elapsed time — live counter updated every second, shown only for running jobs
  • “Queued” label — shown for pending jobs waiting for an available agent

For Coding Agent tasks, a short task ID (#xxxxxxxx) appears next to the label so you can cross-reference with the Task Queue or Task Plans view.

Running jobs are always sorted to the top within each project group. When both running and pending jobs are present at the same time, a footer line shows how many are queued.

What’s normal — expected durations

JobTypical duration
Processing Conversation (Architect)5–20 seconds
Generating Plan2–4 minutes
Generating / Iterating Milestones30–90 seconds
Generating Tasks30–60 seconds
Executing Code (Coding Agent task)1–10 minutes per task
Validating Gate (Deploy)30 seconds–5 minutes per gate
Full deployment10–25 minutes

Jobs that exceed 2× their typical duration may be stalled — see the troubleshooting section below.

What to do when a job fails

  1. Open the job dropdown and note which job failed and in which project.
  2. Navigate to that project and mode — the relevant panel will show an error message or a failed status.
  3. Read the error detail — each failed job has an expandable error message in its panel.
  4. Act based on the error type:
Error typeAction
LLM timeoutRetry the operation — transient timeouts usually resolve automatically
Git push rejectedCheck if someone pushed to the branch concurrently; rebase and retry
Insufficient creditsTop up your balance via Credits & Billing and retry
Configuration errorFix the config (wrong token, invalid repo URL) and retry
Coding agent compilation failureRead the task error, refine the task description, and requeue
  1. Retry — most operations have a retry or re-run button in their respective panel.

Health indicator

Alongside the Active Jobs indicator, the sidebar footer shows a Health indicator — a dot that reflects connection status to the Ileen backend:

StateMeaning
🟢 ConnectedBackend is reachable; all operations nominal
🟡 CheckingPolling in progress
🔴 DisconnectedCannot reach the backend — check your network connection

If the indicator shows Disconnected while jobs are supposedly running, wait a moment. If it doesn’t recover, refresh the page. Running jobs are not lost — they continue on the backend and results will appear once the connection is restored.

Next steps