On this page
Manage collection
Choose collection programs, configure them, target agents, and verify that a deployment produces usable measurements.
Use Manage → Programs, Configurations, and Deployments to answer three separate questions: what should be measured, with which options, and on which devices and schedule. A successful setup ends with a recent observation from the intended agent. A green agent status alone confirms only that the device can contact Lens.
Start with the question and expected evidence
Write down the question before selecting a program. For a WiFi pilot, that might be “Can we collect signal and connection observations on the lobby floor during the morning route?” Identify the location, the devices, the time window, and the fields you expect to see. This makes the first run verifiable and prevents a technically valid deployment from producing irrelevant data.
| Decision | Where to check it | Evidence of success |
|---|---|---|
| Collection work and outputs | Manage → Programs | The program supports the intended device and declares the measurements you need |
| Program options | Manage → Configurations | The selected option set matches the pilot and can be explained to the operator |
| Target devices and timing | Manage → Deployments | The expected agents receive the assignment for the chosen schedule |
| Agent contact and execution | Deploy → Agents | A recent contact, synchronized assignment, and a completed run |
| Result | Metrics view or the relevant solution report | New observations from the correct agent, program, and period |
Choose a program
Open Manage → Programs and inspect the available work and its declared outputs. Check the supported agent type, required device capabilities, and what each output means. A program that can run on Android is not automatically suitable for a desktop host. An output named “signal” and an output named “latency” also answer different questions; confirm units and context before comparing values.
For the MareaClara WiFi example, choose a program that can collect the observations needed for Mobility, such as position and WiFi measurements. The program definition does not select a hotel, choose devices, or set a schedule. Those are later decisions. If the available outputs do not answer the pilot question, resolve that gap before creating a deployment.
Set the program options
In Manage → Configurations, select or prepare a named set of options for the program. The options determine how the work runs and can affect the meaning of the resulting measurements. Keep the device population and schedule out of this decision; they belong to a deployment.
Document the pilot configuration well enough that another operator can repeat the run. Check any required inputs with the person who owns the measurement task. Before editing a configuration used elsewhere, identify the deployments that reference it. A shared change may affect more than the pilot, even when its name stays the same.
Target a small pilot
A deployment combines the program and configuration with agent-selection tags, a schedule, and a timeout. In Deploy → Agents, confirm that the pilot devices have the intended tags and location. Then use Manage → Deployments to assign the work. Review the target population before saving: a broad tag may include more devices than expected, and the collection frequency may affect device battery or data use.
For MareaClara, start with one prepared Android device at one property and a route through a known floor. Check that the chosen tags match that agent. A later rollout can use separate deployments for other properties or device groups while keeping the measurement purpose consistent. Do not assume that matching tags prove an agent is ready: it must also be installed, connected, and able to use the required device capabilities.
Verify the run, then expand
- In Deploy → Agents, check the device’s last contact and whether it received the intended assignment. Use the agent’s Audit view if synchronization or execution failed.
- Wait for the scheduled run. Inspect metrics for the expected program, agent, and time period. Check timestamps, fields, and units rather than only the count.
- For a Mobility pilot, confirm that the measurements have the location and spatial context needed for the analysis view. A raw WiFi reading without usable position cannot establish coverage for an area.
- Compare the result with the original question. If it is suitable, record the program, configuration, tags, schedule, and an example observation before widening the rollout.
| Symptom | Likely next check |
|---|---|
| Agent does not contact Lens | Installation, device connectivity, and agent status in Setup and agents |
| Agent contacts Lens but has no work | Tags, deployment target, and synchronization |
| Work is assigned but no metric arrives | Schedule, program requirements, device permissions, execution audit, and selected time range |
| Metrics arrive but the report is empty | Solution location, period, filters, and whether the readings have the required context |
Control changes and access
A new program version or changed output meaning can make old and new observations hard to compare. Review affected configurations and deployments, test on one agent, and verify a new metric before extending the change. Record when the change occurred so analysts can interpret trends across it.
In Manage → Users, check that each operator has the access needed for their task and location. Program and deployment administration, agent registration, and analysis access are distinct responsibilities. If a user cannot see a property, check their status and location access before concluding that the property’s data is missing.
Continue with Lens concepts for the model behind these decisions, or Mobility collection for a WiFi field example.