A plain definition for EUC teams: application performance monitoring that times the real user GUI for Citrix, Microsoft RDS, and cloud sessions.
Direct answer: On-the-glass APM is application performance monitoring that measures response time and availability from the user’s screen - login-to-shell, published-app launch, and in-session steps - using the same visual path a person takes. For virtual apps and desktops, that means the HDX/RDP session GUI, not only protocol counters, portal HTTP, or host CPU. CitraTest APM® implements on-the-glass APM with synthetic image-recognition transactions and no agents on production session hosts.
Why “the glass” matters in EUC
Citrix and RDS stack many layers: broker, gateway, profile, FSLogix, image, app, and network. Infra can look healthy while the published app still paints late or hangs on a dialog. On-the-glass APM closes that gap by timing what appears on the desktop - the same signal that drives help-desk tickets.
- User-centric - SLA on usable UX, not only session connected
- Works when apps are thick / published - not limited to browser RUM
- Complements observability - Dynatrace/New Relic/Datadog-class tools instrument; glass measures the session UX they often miss for EUC
On-the-glass vs other monitoring styles
| Style | What it sees | Gap for Citrix/RDS |
|---|
| Host / infra | CPU, memory, disk | Green hosts ≠ usable glass |
| Protocol (ICA/HDX, RDP) | Session and network signals | Can miss slow on-screen apps |
| Full-stack APM / observability | Traces, logs, code, infra | Often weak on published GUI path |
| On-the-glass APM | Screen-ready timings via synthetics | - |
Related: What is application monitoring?, APM based on real user experience, HDX vs real user experience, Citrix & RDS APM. Measure response time from the glass: Citrix, Citrix stack,
RDS, cloud VDI, Epic, Oracle Health (Cerner), Remedly, Deloitte, Wave, Paylocity, Epicor, MRPeasy.
How on-the-glass APM typically works
- A robot (synthetic) logs on like a real user - Citrix Workspace, RDP, or cloud client.
- It drives keyboard/mouse through a business transaction (for example open Epic Hyperspace, search a patient, open a chart).
- Image recognition (or equivalent visual checks) waits until the expected screen state appears and records the time.
- Results feed SLAs, alerts, and trends - around the clock, not only during a load test.
Tevron’s product for continuous on-the-glass monitoring is CitraTest APM. For capacity and change validation under concurrency, see CitraTest VU.
When to use it
- Production watch after go-live (morning logon storms, mid-day lag)
- SLA proof for business owners of published apps
- EHR / Epic, Oracle Health (Cerner), Remedly, Sage, Epicor, MRPeasy, and other thick clients on Citrix, AVD, or RDS - Epic on-the-glass
- Whenever Director is green but users still complain
FAQ
What does on-the-glass mean?
Measuring performance from what appears on the user's screen (the glass) - pixels, dialogs, and ready states - rather than only ICA/HDX counters, HTTP, or host CPU.
Is on-the-glass APM the same as full-stack observability?
No. Observability platforms instrument code, traces, and infrastructure. On-the-glass APM answers whether the published app or desktop is usable from the user's viewpoint - often the gap left for Citrix and RDS.
How is it different from Citrix Director?
Director is essential control-plane and session telemetry. It does not time published-app paint the way a person experiences it on the glass.
Does Tevron require agents on VDAs or RDSH?
No. CitraTest APM runs synthetics from outside production, like real clients.
Is on-the-glass only for monitoring, or also load testing?
The same idea applies to both. CitraTest APM watches production continuously; CitraTest VU load tests capacity from the glass before and during changes.
CitraTest APM® and CitraTest VU® are Tevron products. On-the-glass describes the measurement approach, not a third product SKU.