Programmable edge data capture

Run the collector
where the data lives.

Lens manages lightweight agents that execute configured workloads beside devices and local systems. Its control plane defines what runs, where and when, then receives typed output and operating state.

Control plane

Trazadera Lens

ProgramsConfigurationsDeploymentsFleet state
Commands and synchronizationOutput and runtime state

Edge runtime

Lens Agent

ConnectorLocal logicTyped output
LinuxWindowsmacOSAndroid

A controlled deployment path

Keep collection logic separate from the place it runs.

Lens turns an edge requirement into an explicit operating contract. Define the workload and its output, configure the local environment, select the agents, schedule execution and keep the resulting fleet visible from one control plane.

01

Define

Package the collection logic and declare the typed output the program is expected to return.

Executable workload
02

Configure

Keep endpoints, paths, register maps and environment parameters outside the program.

Reusable configuration
03

Target

Use tags and locations to assign each workload to the agents that should receive it.

Explicit scope
04

Execute

Run the workload beside the source on a defined schedule and within an explicit timeout.

Local processing
05

Operate

Inspect heartbeat, synchronization, system information, logs and deployment state centrally.

Fleet control

Northstar Plant

One control plane. Four technical surfaces.

This isolated industrial fixture uses the current product to model an edge estate. Fleet state, executable collection, environment-specific parameters and deployment policy remain separate, visible and independently manageable.

Observe

Keep the edge estate visible.

Each runtime exposes its platform, location, assigned tags and synchronization state before a workload is allowed to run.

Lens fleet view showing Linux, Windows, macOS and Android edge runtimes across an industrial estate
01 / 04

Lens on Android

Set up on site.
Stay connected.

Pair an Android device with Lens, check its connection and inspect runtime logs from the same app.

Know the agent is online.

The device shows its identity, version and connection state. Open the console when you need to look further.

Device-side runtime
Lens Android app showing agent identity, version, Online status and the Agent Console button
ConnectionClick screen to enlarge

Technology

A programmable runtime with a central control plane.

Lens combines a Java and Spring Boot service, durable messaging and portable edge runtimes. The architecture keeps workload definition, delivery, local execution and operational evidence as distinct technical concerns.

01

Java control plane

The current core runs on Java 21 and Spring Boot, exposing the management model through a REST API and web interface.

02

Portable agent

A Java 8-compatible runtime supports Linux, Windows and macOS, with a dedicated Android implementation for mobile edge devices.

03

Durable command transport

Apache Artemis and MQTT/JMS carry instructions and telemetry between the central service and intermittently connected runtimes.

04

Declared outputs

Programs state the fields and types they emit, making collected results explicit before a deployment is scheduled.

05

Decoupled configuration

Environment parameters remain separate from executable logic so programs can move across sites without embedding local detail.

06

Tag-based deployment

Programs, configurations, schedules, timeouts and agent tags combine into one controllable execution contract.

07

Runtime observability

Heartbeat, synchronization, system information and logs make the operating condition of each agent available centrally.

08

Persistent management state

PostgreSQL retains the platform model while APIs keep its resources available to automation and adjacent systems.

Practical scope

Start with one inaccessible source.

A useful first scope identifies the data that cannot be collected reliably from a central system, the environment beside it and the smallest output contract that makes that data usable.

Discuss an edge source

Runtime estate

Deploy close to devices and local systems.

LINLinuxWINWindowsMACmacOSANDAndroid

The supported runtime depends on the agent package and connector required for the source.

Collection contract

Define the connector, execution and output together.

Devices and sensorsMachines and OT systemsLocal APIs and servicesFiles and operational logsNetwork equipmentProprietary local systems

A source is included when a suitable connector already exists or is implemented for the engagement.

Start with the source

Which operational data never reaches a central system?

Bring one source, the environment beside it and the decision its output should support. We can define the connector, execution location, cadence and output contract before discussing a wider estate.

Contact Trazadera