A morning login rush and a full-day concurrent-user sustain answer different questions. Run them separately.
Direct answer: A logon storm is a steep concurrent ramp that stresses auth, profile attach, and first paint. Density / sustain holds target concurrent users and asks whether the session stays usable. Mixing them in one ramp hides whether you failed on logon or on mid-day UX. Measure both on the glass with CitraTest VU® (capacity) and keep watching with CitraTest APM® after go-live.
Side-by-side
| Logon storm | Density / sustain |
|---|
| Shape | Steep concurrent ramp | Hold N users for a stretch |
| Proves | Auth, profile, first shell/app paint | Hosts stay usable at target concurrency |
| Typical fail | Slow/failed logons, profile IOPS spike | CPU/memory pressure, app lag under load |
| Metric focus | p95/p99 logon, fail rate | p95/p99 in-session UX, host headroom |
Why a green dashboard is not enough
Director, Azure, or broker health can look fine while login-to-shell or first paint of the published app is late. On-the-glass timing is what users feel—and what a CAB can defend. Same split applies to Citrix, Microsoft RDS, and cloud VDI/DaaS.
How to run the split
- Freeze the image and connection path (Gateway vs lab shortcut).
- Script the real client GUI (StoreFront/RD Web → session → app).
- Baseline one user on the glass.
- Run the storm; record tails and fails; stop.
- Run the sustain at target density; watch UX and hosts.
- Stop at a concurrent-user line both tests support.
- Re-test after image, GPO, FSLogix/profile, or autoscale changes.
Where Tevron fits
FAQ
What is a logon storm test?
A steep concurrent login ramp that stresses the path to a usable desktop/app, separate from holding many sessions all day.
What is density testing?
A sustain at target concurrent users that checks whether session hosts and apps stay usable—not only whether sessions stay connected.
Why not one combined ramp?
A single curve blends failure modes. You cannot tell if the farm failed on morning logon or on mid-day density.
Does this apply to AVD / DaaS?
Yes. Autoscale can hide density problems; still run storm and density as distinct tests from the glass.