Skip to content

Dashboard Capabilities

This page maps the current AsyncMQ dashboard behavior against Flower-style operational expectations.

Capability Matrix

Capability Status What you get today
Queue visibility (waiting/active/delayed/failed/completed) Supported Overview + queue pages + SSE updates
Worker list and heartbeat visibility Supported /workers with heartbeat timestamps and pagination
Queue pause/resume control Supported Queue overview and queue detail forms
Job listing with filtering/search Supported State tabs + full-text payload search + task/job-id filters + sorting
Job retry/remove/cancel controls Supported Single-action endpoint and bulk form flow
DLQ inspection and retry/remove Supported Queue-scoped DLQ page and actions
Repeatable definition pause/resume/remove/create Supported Repeatables list + new repeatable page
Live dashboard updates Supported SSE stream (/events)
Action audit trail Supported /audit page with filtering and text search
Metrics trend history in dashboard Supported /metrics page + /metrics/history JSON endpoint
Remote worker process control (start/stop/scale) Not available Use CLI + process supervisor/orchestrator
Multi-cluster federation view Not available One configured backend context per deployment
Built-in fine-grained RBAC Partial Pluggable auth backend; app controls authorization model

Flow Coverage Snapshot

flowchart LR
    A["Queue backlog"] --> B["Inspect queues/jobs"]
    B --> C["Act: retry/remove/cancel/pause"]
    C --> D["Audit verification"]
    D --> E["Metrics and worker trend checks"]

Backend Notes

The dashboard follows the configured AsyncMQ backend and uses the backend contract (list_jobs, retry_job, queue_stats, etc.).

Practical implications:

  • Data freshness and ordering depend on backend behavior.
  • Metrics history is process-local memory for dashboard snapshots, not a long-term analytics database.
  • Cross-process dashboard instances do not automatically merge history/audit memory.

What This Means Operationally

AsyncMQ dashboard is production-usable for:

  • queue triage
  • failed-job recovery
  • worker visibility
  • operator action traceability
  • short-window trend checks

It is not a full worker fleet control plane.

For process lifecycle and infra control, use:

  • CLI (asyncmq worker ...)
  • your orchestrator (systemd, Kubernetes, Nomad, ECS, etc.)
  • external monitoring/logging systems for long-term storage