A multi-hour demo is a real load
The map refreshes its data cyclically, the work feed is polled every 3 s, and every additional browser tab adds to the load. When a demo runs for several hours, with alarms and repeated scenes, the demo environment behaves much like a small production deployment.
The practical conclusion: size the demo server with a resource margin and monitor it during the demo instead of treating it as a toy environment. This way the board sees the system, not its limits.
Sizing with a margin
No single number fits every demo, so size based on a trial. Run the full scenario on the target machine, open the map in several tabs, trigger the alarms and watch resource usage for at least one hour.
Add a margin to the result that will carry the peaks, for example when the board asks to repeat a scene. Spare memory is cheaper than a failed demo.
It is good practice to run the trial under conditions close to the demo: the same number of map tabs, the same demo accounts and the same network link the room uses. The gap between a trial on a laptop and a demo in the room is often larger than expected, especially when additional viewers join the network.
- a trial of the full scenario on the target machine,
- several map tabs open at the same time,
- resource usage observed for at least one hour,
- headroom for repeated scenes and extra viewers.
Timing parameters of the demo server
Monitoring during the demo
During the demo one person, ideally the IT administrator, watches resource usage and the state of the services. Running the demo itself takes up the presenter's attention and leaves no time to check the state of the machine.
It is worth setting up simple alerts: a threshold for resource usage and a lack of response from the service. A quick reaction lets you refresh the service before the board notices a problem. It is also good to have a fallback plan: a recording of a walkthrough of the system as spare demo material.
Agree in advance with the presenter on the signal after which the administrator steps in, and on the place from which they monitor the environment. This simple split of roles shortens the reaction time and takes the pressure off the person leading the demo.
Start, stop and status with one script
The demo starts and stops with the start or stop command from a single script, and its state can be checked with the status command. This approach reduces the number of manual steps, and every manual operation is an opportunity for an error under time pressure.
Stopping the demo should leave the instance in a known state: events closed, no active demo alarms, and the demo's own markers removed.
Test the script before every demo, even when nothing in the environment has changed. The output of the status command should clearly show whether the services are running, whether the demo alarms are switched off and whether the demo accounts are ready to use.
Markers of people and their cleanup
A person's marker on the map lives in the cache for about 780 s, and the demo removes only the markers of its own demo accounts. This safeguard protects the data of other users of the instance.
For a demo longer than two hours, stopping may not remove every marker, for example when the connection to the cache was closed in the meantime. Include a check step in the stop procedure and inspect the map once the demo ends.
Good practice is a short checklist after the demo: is the map clean, do the demo accounts have no active sessions, and has the password been changed.
If the demo is repeated over several days in a row, introduce a fixed order: check the state before the demo, stop and inspect the map after it, and restart the next day from a clean state. This routine removes surprises caused by leftovers from the previous demo.
Video
Frequently asked questions
How do I prepare the server for a multi-hour demo?
Size it based on a trial of the full scenario, with several map tabs open, and add a resource margin for repeated scenes. Monitor resource usage during the demo.
How do I stop the demo?
With the stop command from a single script, and check the state with the status command. Stopping removes the markers of the demo's own accounts.
What should I do with the markers after a long demo?
After a demo longer than two hours, check the map and, if needed, remove the remaining markers of the demo's own accounts.
A demo server without surprises
We will help you size and test the demo environment before the demo for your board.
Let's talk