Skip to content

SiloSentry #

SiloSentry sells grain-level monitors. Its customers need to plan grain movements and know whether a reading can be trusted. Its small engineering and support team needs to build the product, hand over each installation and investigate what happens after shipment.

This is a fictional business and a working reference application. The captures below come from the current Dataplicity Applications workspace and Customer Portal, Files, Messaging, incident response and status pages, using fictional team records, simulated equipment, sensor observations and heartbeats. They show software behaviour, not a real customer deployment or physical sensor qualification. Captures were refreshed on 27 October 2026.

Light-mode Ashton Mill portal showing North Silo at 68.2 of 120 cubic metres and South Silo below its threshold
The customer sees calibrated quantities, reading freshness and history. South Silo’s separate low-level condition remains visible.

The full business story lives at silosentry.com/inside.html. For the implementation path, use Build a customer-operated level monitor.

Keep one installation in view #

ItemReference example
Customer and siteAshton Mill
AssetNorth Silo
MonitorSS-AM-01
Device ClassSS Level Monitor
ApplicationSiloSentry
Capacity120 m³
Low threshold25 m³
Normal readingApproximately 68.2 m³
Raw streamlevel, measured height in millimetres
ComparisonSouth Silo / SS-AM-02: 18.5 of 80 m³, below its 20 m³ threshold

SiloSentry supplies the hardware, sensor adapter, physical calibration, commissioning and customer support procedure. Dataplicity supplies the device connection and identity, reusable application and portal components, customer scope, history, alerts and authorised operational tools.

Alex defines the product #

Open Engineering → Applications and open SiloSentry. The current workspace has Overview, Customer app, Automations, Data, and Test & publish. Older screenshots calling these Build, Access, Behaviour and Release use the earlier vocabulary.

Under Data, the application connects its sensor Device Class and defines:

  • a Silo business object with the sensor, input, capacity, units and thresholds;
  • Strapping row records that relate each silo’s measured height to its volume;
  • read-only current height and volume derived from the configured device stream.

The reference’s level stream is in millimetres. A publisher reporting metres must convert its values before using this contract. A channel label does not perform that conversion.

Configure reusable customer screens #

In Customer app → Workspace pages, the application uses the shared Capacity layout bound to Silo data. Each record receives the generic capacity, history and optional camera presentation configured by the application. The same product widgets are used when opening a silo and on its monitor’s Controls view.

SiloSentry Customer app workspace with eight configured pages and process presentation and role sections
The current Customer app workspace owns customer pages, process presentation and roles.
Reusable Capacity page editor with Silo selected as the business data
Bind a reusable page to business data. The editor’s small preview uses illustrative values; the running customer records are separate.

The application also publishes Work and Exceptions pages. An optional camera is shown only when it is actually assigned; this example does not need a blank camera or signage panel.

Review the threshold and stale-data behaviour under Automations. Use Test & publish to validate the draft, publish an immutable version and manage its portal attachment and customer instances. A draft edit does not silently change the version a customer is running.

Sam commissions North Silo #

Sam checks the monitor’s serial, assigns it to Ashton Mill and places it on the correct customer network. He then binds SS-AM-01 to North Silo and chooses the correct sensor input.

For this illustrative vessel, the calibration points are:

Height (mm)Volume (m³)
00
2,50030
5,00060
10,000120

A reading near 5,683.3 mm therefore gives approximately 68.2 m³. This table is a software example, not a validated shape or calibration for a real grain silo. At least two valid points are needed to show calibrated volume history. For a real installation, use a validated strapping artefact for that vessel.

North Silo sensor binding, capacity, thresholds and height-to-volume calibration table
Match the physical vessel, sensor and calibration before relying on a customer quantity.

The physical installation still needs measurement checks, power/network tests, mounting and replacement procedures. Storing installation notes or a drawing supports the handover; it does not perform those checks.

Keep the installation context with the team #

Alex records the revision B interface baseline and bench-check plan in Files → Ashton Mill → North Silo - SS-AM-01. Sam keeps Commissioning and support.md and North Silo calibration.csv beside it. The handover record names the monitor, symptom, completed checks, verified software result and remaining owners. Physical checks still need real results.

Actual Files screen showing the three shared North Silo SS-AM-01 reference documents
One internal installation folder carries the engineering baseline, calibration and support record.

In Messaging, open Ashton Mill - SS-AM-01 handover. Sam records that the gateway is online while the North reading is stale; Alex takes the startup investigation; Jo can give the customer an accurate update. The conversation links directly to the shared folder. File references do not grant extra access.

Actual Messaging conversation with Alex, Sam and Jo handing over the North Silo investigation
The authored example runs in real Chat: reviewed context, completed checks and the next owner.

These are fictional records seeded into an isolated local organisation through the normal Files API and team-chat services. Documents can be previewed and downloaded; messages are written by the demo fixture. No AI response, physical bench result or real customer correspondence is fabricated as evidence. The mill's portal does not grant access to the internal team folder or conversation.

Jo checks the customer experience #

Jo prepares the handover and verifies the mill’s view. The operator can compare North and South Silo, inspect history and identify a low-level condition without needing the OEM’s engineering workspace.

Open a silo for its current reading, history, bounded configuration and calibration. Open its monitor from Devices for the device’s own Controls and Diagnostics. The sensor reading and connection state answer different questions.

For daytime use, choose Account → Appearance → Light. Dark mode remains available. The preference is saved for that browser; charts, forms, notices and status colours follow the chosen theme.

Before handover, test the intended viewer, operator and administrator roles, then sign in as another customer and confirm that Ashton Mill’s equipment and records are absent. These access checks are required product qualification; the screenshots alone are not proof of isolation.

Distinguish old data from a disconnected monitor #

In the stale scenario, SS-AM-01 continues to heartbeat but its sensor observation is ten minutes old. The portal retains 68.2 m³ as the last known quantity and marks it Stale. It does not turn an old observation into an empty silo or a fresh healthy result.

North Silo marked stale while the comparison silo has fresh observations
A connected device can still report stale product data.

The offline scenario also ages the simulated heartbeat. The portal reports the sensor offline and retains its last known volume.

North Silo sensor offline state retaining its last known volume and history
Connection loss is a separate condition from sensor freshness or a low grain level.

In a real installation, Sam would investigate the sensor adapter, service logs, cabling and installation record, using authorised remote tools where needed. This reference run resumes the simulator instead of claiming a remotely repaired physical machine.

Watch the fleet and own the response #

Use a fleet monitor for a selected connectivity cohort. The reference configures Ashton Mill monitor connectivity in the customer's fleet organisation, using Any offline, three failing evaluations to trigger and three successes to recover. The local reference fleet uses the same Business/beta configuration as its vendor; access follows the current plan and release-channel policy. Preview its two members. Connectivity and sensor freshness are different: an online gateway does not establish that its sensor is producing new data.

Device and fleet connectivity monitors create alerts. Service, heartbeat and user-impact monitoring can feed incident automation when configured. This worked sensor incident is opened manually; it does not demonstrate automatic escalation from a product freshness alert.

When a response needs ownership, open the incident Delayed readings on one installation. Sam acknowledges it, follows the checks in the handover, and records the recovery. Configured on-call schedules can escalate an unacknowledged incident; the isolated example has no escalation path or external recipients.

Actual incident record with Sam as responder and the acknowledgement and recovery history
An incident gives the response an owner and a history separate from the discussion.

Give customers a selected status update #

Jo reviews the customer-facing wording and selects SiloSentry service updates. The local status page shows the affected service, current action and recovery update. Keep device identifiers, diagnostic material and internal discussion out of public summaries and timeline comments. Publishing an incident is an explicit choice; the example does not automatically publish every alert.

Local SiloSentry status page showing selected customer-facing incident updates
The status page communicates selected updates. The reference reporting-process heartbeat uses a real local request; it does not claim production fleet uptime.

Monitoring, incident response, status pages, shared files and team discussion are parts of the same support process. Handling them together can reduce the separate services the equipment business buys and integrates. Configure scopes, response policy and communication for your product; availability and limits follow the organisation's Dataplicity plan.

Verify recovery from the product signal #

Recovery sends a fresh sensor observation and current heartbeat through the normal data services. North Silo returns to a live reading near 68.2 m³, and its stale alert resolves. South Silo’s independent low-level alert stays open.

Exceptions page retaining the separate low-level alert after North Silo recovers
Clearing one problem does not clear an unrelated operating condition.

A successful command alone would not establish this result. For real equipment, also compare the recovered measurement with a trusted site reference and record what was done, what was checked and any follow-up work.

What this reference proves #

The current example exercises a published application, customer records, height-to-volume conversion, reusable portal widgets, history, low/stale states, heartbeat-based connectivity and recovery. Light-mode captures come from the actual portal, not recoloured screenshots.

It does not establish physical accuracy, production field qualification, completed commissioning, a real customer result or a repaired machine. Extend the same checks to your own hardware, roles, customer boundaries and deployment before making those claims.