Activity Monitoring — Tasks and Journal
Most databases run your queries and imports in the dark. You fire off a large load, the cursor spins, and you wait — with no idea whether it’s making progress, where it’s running, or why it failed.
WWKG Studio gives every workspace two live, side-by-side panels in the sidebar: a Tasks panel showing what’s happening now, and a Journal panel showing what has happened. Both are scoped to the workspace you’re working in, and both update live as operations run.
The Tasks panel — what’s happening now
The Tasks panel lists the workspace’s in-flight work — queries, imports, commits, branch operations — and keeps each one visible for a short while after it finishes, so quick operations don’t flash past before you see them.
For each task you get:
- Live progress — a progress bar and phase for running work, plus a memory footprint for the heavy operations.
- Cancel — stop a running task you started, right from the row.
- Active vs. recently finished — live work sits up top; just- finished work collapses into a dimmed tail you can hide.
- Failures stand out — a failed task stays at full strength with a red accent and sorts ahead of the successes, so an error never quietly scrolls away.
Sort and kind chips let you focus the list, and the top-bar Tasks icon carries a count badge — red when there are recent failures to look at — so you know there’s something worth opening even when the panel is closed.
The Journal panel — what has happened
The Journal panel is the workspace’s history: one card per task, in order, with its outcome, timing, and the events it went through. Expand a card to see the per-event detail. Status and kind chips filter the view; failed cards are tinted red so problems are easy to spot in the record.
The journal travels with the workspace — it’s encrypted with your workspace key and replicated like any other workspace data — so the history is available wherever the workspace is, and confidential at rest.
Provenance — who, where, and across nodes
Both panels show the full execution path of a task: who initiated
it, the node it entered through, and the node it actually ran on. When
work is delegated to another node — for example a SERVICE query or a
remote data fetch handed to a worker — the task is labelled with that
node’s role (host, worker, custodian) and nests under the operation
that spawned it, so a distributed operation reads as a single tree
rather than scattered rows.
Operations that act on the workspace itself but can’t be recorded inside it — deleting a workspace, say — are surfaced too, marked so you can tell host-level work apart from your data operations.
Why it’s built in
Because activity is first-class in WWKG — every operation is a tracked task and every state change is a journal entry — you get this observability without wiring up a separate metrics stack, log aggregator, or tracing backend. It’s the same record the system itself uses to manage and recover work, surfaced for you in Studio.