Tasks
Inspecting and stopping individual running task replicas.
A task is one running replica of a deployment. If a deployment has replicas: 3, there are three tasks. Each has its own state and log stream.
State lifecycle: pending → starting → running → stopping → stopped | failed
GET /api/v1/tasks — List all
Lists tasks across all deployments and apps on the platform.
GET /api/v1/tasks/ — Get task
Fetch a single task and its current state.
curl http://localhost:47080/api/v1/tasks/$TASK_ID{
"id": "uuid",
"deployment_id": "uuid",
"node_id": "uuid",
"state": "running",
"published_port": 48123,
"created_at": "...",
"updated_at": "..."
}POST /api/v1/tasks//stop — Stop
Requests a graceful stop. The agent stops the task; the scheduler may reschedule it depending on the deployment's replica count.
curl -X POST http://localhost:47080/api/v1/tasks/$TASK_ID/stopGET /api/v1/tasks//logs — Stored logs
Fetch stored log entries for this task.
curl http://localhost:47080/api/v1/tasks/$TASK_ID/logs | jq '.[] | .message'POST /api/v1/tasks//logs — Append logs
Used by zectre-agent to write log entries. Not typically called directly.
GET /api/v1/deployments//tasks
List all tasks for a specific deployment — same response shape as /api/v1/tasks but filtered.
curl http://localhost:47080/api/v1/deployments/$DEPLOY_ID/tasksSee also
- Deployments — log streaming, scaling, promotion
- Internal services — scheduler stop-task endpoint used by agents