Documentation
CurrentOp Dashboard
Live view of every running operation on the cluster. Filter by op type, namespace, and running time. Kill a misbehaving op safely — with audit envelope.
What it is
A workspace tab fed by `db.currentOp()` with filters and a kill button. Each row shows opid, namespace, op type (query / update / getmore / command), running time, planSummary, and the full filter / pipeline. Kill emits a `currentop.kill` audit line including the opid and the reason you provided.
When to use
Open during incident response when an app is hanging on a query. Open during routine review to confirm no orphan ops linger. Useful for hunting transactional locks blocking a deploy.
How to open
- In the NoSqlStudio desktop app, open the Tools menu → Dashboards → CurrentOp, or press Ctrl+Alt+Shift+O.
- A workspace tab opens with a live-refreshing table of all running operations.
- Pick a database in the picker — the table filters to that database's ops only.
Working with it
- Filter — op type (query / update / getmore / command), running time (slow > N ms), namespace, planSummary contains "COLLSCAN".
- Click any row to expand the full op JSON — including ns, plan, lockStats, command details.
- Kill op — right-click → Kill operation. The engine prompts for a reason (required), runs db.killOp(opid), and emits a currentop.kill audit line.
- Right-click also offers "Start recording" — pipes this view's samples to the Monitoring Recorder for retrospective analysis.
Limitations (v1)
- killOp is best-effort — server may decide the op cannot be interrupted (e.g. yielded, blocked on a chunk migration).
- Free tier — some currentOp fields hidden by Atlas. The expanded view shows the gaps.
- Hyperactive clusters — at high op rates, a single tick may show different ops than the next; the Recorder gives stable retrospective data.