Skip to content
NoSqlStudio
Documentation

Realtime Dashboard

A single workspace tab giving you the live cluster overview — CPU and SYS MEM with self-correcting formulas, disk pressure, hottest collections, slowest queries. Free-tier detection (M0/M2/M5) shown as a red pill.

What it is

A snapshot view of server activity refreshed at a fixed cadence. Pulls serverStatus + selected $currentOp + $indexStats samples. CPU is approximated from threads and tcmalloc; SYS MEM from cache + resident; disk from log-storage stats. Atlas free tier is detected by the absence of extra_info / tcmalloc / wt-cache.max and surfaces a banner explaining what the dashboard can and cannot show.

When to use

Open during routine cluster check-in or when investigating a "is something wrong?" hunch. Lower latency than the Atlas UI for the most common signals. For sustained monitoring, use mongostat / mongotop or wire alerts via the Alerting webhooks workspace.

How to open

  1. In the NoSqlStudio desktop app, open the Tools menu → Dashboards → Realtime Dashboard, or press Ctrl+Alt+B.
  2. A workspace tab opens with cards for CPU, SYS MEM, Disk, Hottest collections, Slowest queries. Live data starts flowing immediately.
  3. Atlas free tier banner appears at top when extra_info / tcmalloc / wt-cache.max are missing — explains why certain cards show degraded data.

Working with it

  • CPU card — derived from active connection count + thread state + tcmalloc allocations. Capped at 2 vCPUs by default for cloud single-node sanity. Free-tier shows a red "FREE TIER" pill.
  • SYS MEM card — six signals (cache, resident, virtual, page faults, swap, available) combined into a 0–100 MB Atlas estimate. Same free-tier pill applies.
  • Hottest collections — top N by ops/sec via aggregated currentOp. Click a row to jump to the collection in the explorer.
  • Slowest queries — top N by p99 latency. Click to inspect the actual operation including filter, sort, and indexes used.

Limitations (v1)

  • Free tier (M0/M2/M5) — Atlas hides extra_info so CPU/SYS MEM estimates are heuristic only. The banner clarifies which cards are downgraded.
  • Refresh cadence is fixed at ~1s. For longer-term trends use mongostat (which records to a sliding window) or wire Alerting webhooks.
  • The dashboard reads from the active connection only. To compare clusters open multiple connection tabs.