Industrial IoT projects often begin with sensors.

A vibration sensor is installed on a motor.

A power meter is connected to an electrical panel.

A weather station begins sending temperature, humidity, rainfall, and wind data.

A Modbus gateway starts collecting measurements from several field instruments.

At this point, however, the organization does not yet have a useful monitoring system.

It has data.

The next challenge is turning that data into information that engineers, operators, maintenance teams, and managers can actually use.

This is where Grafana becomes valuable.

Grafana provides a flexible visualization and monitoring layer that can connect to multiple data sources, query operational data, transform it, visualize it in dashboards, and use that data for alerting. Grafana’s current documentation emphasizes that data sources remain where they are stored; Grafana queries them rather than requiring every data source to be migrated into Grafana itself.

For Industrial IoT applications, this makes Grafana particularly suitable for systems involving:

environmental monitoring,
machine condition monitoring,
electrical power monitoring,
water and wastewater monitoring,
energy management,
process utilities,
remote equipment,
building monitoring,
and multi-site operations.

A typical architecture might look like:

Sensors

↓

Edge Gateway

↓

InfluxDB / PostgreSQL / Time-Series Database

↓

Grafana

↓

Dashboard + Alerts + Analytics

The dashboard is therefore only one part of the complete Industrial IoT architecture.

This guide explains how to design that architecture and build an effective Industrial IoT dashboard using Grafana.

What Is Grafana?

Grafana is an observability and visualization platform used to query, visualize, and monitor data from multiple data sources.

A Grafana dashboard consists of panels displaying data retrieved from one or more connected sources.

According to current Grafana documentation, each visualization panel is driven by one or more queries to a configured data source.

Grafana can connect to many types of data stores, including:

InfluxDB,
PostgreSQL,
Prometheus,
MySQL,
Elasticsearch,
Loki,
cloud monitoring platforms,
and other supported data sources.

For Industrial IoT, two particularly useful options are:

InfluxDB

and:

PostgreSQL / TimescaleDB

Both can work well, but they serve slightly different architectural purposes.

1. Start with the Industrial Problem

Before opening Grafana and creating charts, define the operational problem.

This is one of the most important steps.

A dashboard should not begin with:

Which visualization should we use?

Instead, begin with:

What decision should this dashboard help someone make?

Examples include:

Machine Condition Monitoring

Is the motor operating normally?
Is vibration increasing?
Is bearing temperature deteriorating?
Which machine requires maintenance first?

Electrical Monitoring

Which feeder consumes the most power?
Is power factor deteriorating?
Did a voltage sag occur?
Which transformer is approaching its loading limit?

Environmental Monitoring

Is PM2.5 above the target level?
What is the current wind direction?
Is rainfall increasing?
Which monitoring station has an abnormal reading?

Water Monitoring

Is the tank approaching high level?
Has pump efficiency decreased?
Is pH outside the operating range?
Which pumping station is consuming excessive energy?

These questions determine which measurements belong on the dashboard.

2. Define the Complete IoT Architecture

A professional Grafana deployment should not be treated as an isolated visualization project.

The entire data path must be designed.

A typical Industrial IoT architecture contains five layers.

Layer 1 — Sensors and Field Devices

Examples include:

vibration sensors,
temperature sensors,
humidity sensors,
pressure transmitters,
flow meters,
energy meters,
environmental sensors,
weather stations,
PLCs,
protection relays,
and machine controllers.

Layer 2 — Field Communication

Common protocols include:

RS485 Modbus RTU,
Modbus TCP,
OPC UA,
MQTT,
Ethernet,
CAN,
4–20 mA,
pulse input,
and digital I/O.

Layer 3 — Edge Gateway

The edge gateway collects data from field devices.

Functions may include:

polling Modbus devices,
protocol conversion,
timestamping,
local calculations,
buffering,
data validation,
and transmitting measurements to the server.

Layer 4 — Database

The database stores historical data.

Common choices include:

InfluxDB,
PostgreSQL,
TimescaleDB,
Prometheus,
industrial historian platforms.

Layer 5 — Grafana

Grafana provides:

visualization,
dashboards,
filtering,
trends,
analytics,
alerts,
and user access.

Understanding these layers prevents a common mistake:

Grafana is not the sensor database itself.

It usually visualizes and analyzes data stored elsewhere.

3. Choose the Right Database

Database selection affects the entire monitoring architecture.

For many Industrial IoT systems, two strong options are InfluxDB and PostgreSQL with TimescaleDB.

InfluxDB for Industrial IoT

InfluxDB is designed specifically for time-series data.

Grafana’s current documentation identifies InfluxDB as particularly suitable for high-volume metrics, IoT sensors, industrial monitoring, and high-frequency measurements.

Typical Industrial IoT data has the form:

Timestamp + Asset + Parameter + Value

For example:

2026-09-10 08:00:01
Motor_01
Vibration_DE
3.4 mm/s

One second later:

2026-09-10 08:00:02
Motor_01
Vibration_DE
3.5 mm/s

This is naturally time-series data.

Grafana currently supports InfluxDB OSS, Enterprise, and Cloud variants and can work with SQL, InfluxQL, or Flux depending on the InfluxDB version. The current Grafana InfluxDB data-source documentation also lists alerting and annotations among supported capabilities.

PostgreSQL and TimescaleDB

PostgreSQL is useful when IoT data must be combined with relational information.

Examples include:

customer information,
site hierarchy,
asset records,
sensor configuration,
alarm configuration,
maintenance history,
user information,
work orders,
and equipment metadata.

TimescaleDB extends PostgreSQL with time-series capabilities.

This makes it possible to maintain both:

Operational time-series data

and:

relational business data

inside a PostgreSQL-based architecture.

Grafana includes native PostgreSQL connectivity. Current Grafana documentation recommends that database credentials used by Grafana have only the minimum necessary permissions—typically SELECT access to the relevant tables and schemas—because Grafana itself does not validate whether arbitrary SQL queries are safe.

That is an important security consideration for industrial deployments.

4. Example Siteplore Architecture

A practical architecture for an industrial monitoring platform could look like:

Industrial Sensors

↓

RS485 Modbus RTU

↓

RekaSense Edge Gateway

↓

4G / Ethernet

↓

Siteplore Server

↓

PostgreSQL / TimescaleDB

↓

Grafana Cloud

This architecture works well where the organization wants control over its own database.

Another option is:

Sensor

↓

Edge Gateway

↓

InfluxDB Cloud

↓

Grafana Cloud

The best architecture depends on:

data volume,
cybersecurity requirements,
internet reliability,
retention requirements,
operating cost,
and customer architecture.

5. Design the Data Model Before Designing the Dashboard

Poor data structures eventually create poor dashboards.

An industrial data model should allow measurements to be grouped logically.

A useful asset hierarchy may be:

Customer

↓

Site

↓

Area

↓

Equipment

↓

Sensor

↓

Parameter

For example:

Customer: ABC Petrochemical
Site: Plant 1
Area: Utility
Equipment: Cooling Water Pump P-101
Sensor: Vibration Sensor VS-101
Parameter: RMS Velocity

Another example:

Customer: Mining Company A
Site: Mine 01
Area: Crushing Plant
Equipment: Conveyor CV-101
Sensor: Motor Power Meter
Parameter: Active Power

This structure becomes extremely useful when creating Grafana variables.

6. Use Consistent Tagging

Industrial IoT systems should use consistent metadata.

Useful tags include:

customer,
site,
area,
asset,
equipment_type,
sensor,
parameter,
unit,
location.

For example:

customer=ABC
site=Plant01
area=Utilities
asset=P101
parameter=vibration

Consistent tags allow Grafana dashboards to dynamically filter data.

Without consistent naming, dashboards become difficult to maintain.

7. Connect Grafana to the Data Source

After the database has been prepared, Grafana must be connected to it.

In Grafana Cloud, data sources are configured under:

Connections → Data Sources

For InfluxDB, Grafana’s current guided setup specifies choosing InfluxDB from the data-source list and then configuring connection parameters such as URL, authentication, and query language.

Typical InfluxDB configuration may include:

Server URL
Organization
Bucket
Token
Query language

For PostgreSQL:

Host
Database
User
Password
TLS settings
PostgreSQL version

For production systems, avoid using database administrator credentials.

Use a dedicated read-only Grafana account wherever possible.

8. Test the Data Source

Before building any dashboard, verify the connection.

Use:

Save & Test

and then open:

Explore

Grafana’s InfluxDB integration allows data to be queried in Explore before a dashboard is created.

This is a useful troubleshooting approach.

Check:

timestamps,
units,
asset names,
missing measurements,
sampling frequency,
and query performance.

A dashboard should not be used to debug raw ingestion problems.

Solve those problems first.

9. Build Your First Dashboard

The basic workflow is:

Dashboards

→

New Dashboard

→

Add Visualization

→

Select Data Source

→

Write Query

→

Select Visualization

Each panel normally executes a query against its source and transforms the results into a visualization.

10. Start with an Overview Dashboard

One common mistake is placing every available measurement on the first page.

Instead, build a hierarchy.

For example:

Site Overview

Show only:

overall site status,
active alarms,
critical assets,
total power,
environmental status,
communication status.

Then users can drill down into more detailed dashboards.

11. Example Dashboard Structure

A well-designed Industrial IoT dashboard might follow this layout.

Row 1 — Site Status

Panels:

Equipment online
Equipment alarm
Communication health
Active alerts
Data freshness

Row 2 — Critical KPIs

Examples:

total energy consumption,
maximum vibration,
average environmental condition,
water usage,
production-related KPI.

Row 3 — Trends

Examples:

vibration trend,
motor current,
energy usage,
temperature,
rainfall.

Row 4 — Equipment Detail

Examples:

individual machine condition,
sensor values,
health indicator.

Row 5 — Analytics

Examples:

historical comparison,
anomaly detection,
forecast,
efficiency indicator.

This is easier for users to understand than a page containing twenty unrelated charts.

12. Use the Stat Panel for Current Status

Grafana’s Stat visualization is useful for displaying current values.

Examples:

Motor Vibration

3.2 mm/s

Transformer Loading

68%

Tank Level

74%

PM2.5

22 µg/m³

Power

423 kW

A Stat panel should answer:

What is the current condition?

13. Use Time Series Panels for Trends

Time-series charts are among the most valuable visualizations for industrial engineers.

They reveal:

deterioration,
recurring patterns,
operating cycles,
disturbances,
and relationships between variables.

For example:

Motor Vibration — Last 30 Days

may reveal:

2.1
2.2
2.3
2.6
2.9
3.3
3.7 mm/s

The current value alone might not appear critical.

The trend tells a different story.

14. Combine Related Variables

Avoid creating separate dashboards for every measurement.

Correlated information should be displayed together.

For a pump:

vibration,
bearing temperature,
motor current,
flow,
discharge pressure.

For environmental monitoring:

PM2.5,
wind speed,
wind direction,
rainfall.

For power monitoring:

voltage,
current,
active power,
power factor,
harmonics.

This helps users analyze causes rather than just observe symptoms.

15. Use Gauge Panels Carefully

Gauges are visually attractive but can consume substantial dashboard space.

Use them for measurements where range context matters.

Examples:

transformer load,
tank level,
battery state of charge,
equipment utilization.

Do not use gauges for every measurement.

For trending variables, time-series charts usually provide more information.

16. Use Thresholds

Thresholds help users interpret values quickly.

For example, a motor vibration panel could represent:

Normal
Warning
Alarm

Threshold values should be based on:

engineering limits,
manufacturer recommendations,
operating baseline,
industry standards,
or approved site criteria.

Avoid arbitrary red-yellow-green thresholds.

They should have engineering meaning.

17. Add Units

Always configure engineering units.

Examples:

°C
bar
kPa
mm/s
m³/h
kW
kWh
%
ppm
µg/m³

A value such as:

has little meaning.

But:

45 °C

is immediately understandable.

18. Create Dashboard Variables

Variables are one of Grafana’s most powerful features for Industrial IoT.

Grafana’s current documentation describes variables as a mechanism for creating interactive dashboards that users can filter without editing the underlying queries. Variables also help avoid maintaining many almost-identical dashboards.

Instead of creating:

Motor01 Dashboard
Motor02 Dashboard
Motor03 Dashboard
Motor04 Dashboard

create:

Machine Monitoring Dashboard

with variable:

$asset

Users then select:

Motor01
Motor02
Motor03
Motor04

The same dashboard updates automatically.

19. Recommended Industrial Variables

Useful variables include:

$customer
$site
$area
$asset
$sensor
$parameter

For Siteplore, an effective hierarchy might be:

Customer

↓

Site

↓

System

↓

Asset

This can make a single dashboard reusable across many installations.

20. Example Dynamic Dashboard

Suppose a customer operates 200 pumps.

Without variables:

200 dashboards

would be required.

With variables:

Site = Plant A
Area = Utility
Asset = P-101

The dashboard automatically loads P-101 data.

Change:

Asset = P-102

and every panel refreshes.

This dramatically reduces dashboard maintenance.

21. Use Repeating Panels

Grafana can repeat panels based on variables.

For example, if the user selects:

Pump 1
Pump 2
Pump 3
Pump 4

Grafana can generate a similar visualization for each asset.

This is useful for:

fleet comparison,
multiple motors,
environmental stations,
power meters,
pumping stations.

22. Create an Equipment Health Dashboard

A machine-condition dashboard might include:

Header

Asset:

Cooling Water Pump P-101

Status:

RUNNING

Health:

NORMAL

Current Condition

Vibration DE
Vibration NDE
Bearing temperature
Motor current
Power

Historical Trends

24 hours
7 days
30 days

Operational Context

flow,
discharge pressure,
load.

Alerts

active alarms,
previous alarms,
maintenance events.

This structure supports troubleshooting much better than a generic dashboard.

23. Example PowerWatch Dashboard

An electrical monitoring dashboard might display:

System Overview

Incoming Voltage
Current
Frequency
Active Power
Power Factor
Energy Today

Power Quality

THD Voltage
THD Current
Voltage Imbalance
Current Imbalance

Demand

Current Demand
Peak Demand
Monthly Demand

Historical Trends

Power
Energy
Current
Voltage

Events

Voltage Sag
Voltage Swell
Trip
Communication Loss

The dashboard should allow navigation from:

Site → Substation → Panel → Feeder

24. Example RekaSense Dashboard

An environmental dashboard might display:

Weather

Temperature
Humidity
Rainfall
Wind Speed
Wind Direction
Atmospheric Pressure

Air Quality

PM1
PM2.5
PM10
CO₂
VOC

Environmental Status

Normal
Warning
Alarm

Historical Trends

24 hours
7 days
30 days

Location

Each remote station can be identified by:

site,
latitude,
longitude.

This is useful for mining, agriculture, industrial estates, and environmental monitoring.

25. Example Water Monitoring Dashboard

A water dashboard could contain:

Hydraulic

level,
pressure,
flow.

Water Quality

pH,
conductivity,
turbidity,
dissolved oxygen,
temperature.

Pump Condition

status,
current,
vibration,
bearing temperature.

Energy

kWh/m³ pumped

Alerts

high level,
low pressure,
pump fault,
abnormal water quality.

26. Build Engineering Calculations into the Dashboard

Raw sensor readings are useful.

Derived engineering indicators are often more useful.

For example:

Instead of only showing:

Pump Power = 22 kW

calculate:

Specific Energy = kWh / m³

Instead of:

Plant Energy = 4,500 kWh

calculate:

kWh / tonne production

Instead of:

Compressor Energy

calculate:

kWh / Nm³ compressed air

These normalized KPIs make equipment performance comparable.

27. Data Transformations

Grafana supports transformations that allow query results to be reorganized or combined before visualization. Current InfluxDB documentation explicitly lists transformations among the functions available after connecting the data source.

Transformations can help:

rename fields,
join query results,
calculate values,
organize tables,
filter fields,
and create derived displays.

However, complex calculations are often better performed upstream.

A useful architecture is:

Simple Display Calculation → Grafana

Complex Engineering Calculation → Edge / Database / Analytics Service

This keeps dashboards maintainable.

28. Create Alerts

A monitoring system becomes significantly more useful when users do not need to continuously watch dashboards.

Grafana Alerting can evaluate metrics or logs and identify specified events or conditions. Current Grafana documentation also allows queries and expressions from multiple supported data sources to participate in alert rules.

For example:

Bearing Temperature > 80 °C

could trigger an alert.

But good industrial alerting should include more context.

29. Avoid Simple Alarm Logic Where Possible

Instead of:

Vibration > 5 mm/s

consider:

Vibration > 5 mm/s
FOR 10 minutes
AND Machine Status = Running

This helps prevent alarms when equipment is stopped or temporarily transitioning.

Another example:

Tank Level > 90%
AND rising

is often more useful than:

Tank Level > 90%

30. Separate Alert Rules from Notification Routing

Grafana currently distinguishes between the rule that decides whether a condition is abnormal and the notification mechanisms used to determine who receives the alert. Contact points and notification policies are configured separately.

This is useful for industrial organizations.

For example:

Electrical Alarm

Send to:

Electrical Team

Vibration Alarm

Send to:

Mechanical / Reliability Team

Environmental Alarm

Send to:

HSE Team

Critical Site Alarm

Send to:

Operations Manager

The same monitoring platform can therefore route information according to responsibility.

31. Design Alert Severity

Industrial dashboards should generally classify notifications.

For example:

Information

No immediate action required.

Warning

Investigation recommended.

Alarm

Maintenance or operational response required.

Critical

Immediate response required.

This reduces alarm fatigue.

32. Monitor Data Freshness

One commonly overlooked industrial alarm is:

No Data

A sensor that reports:

Temperature = 28°C

may appear healthy.

But what if that value has not changed for six hours because the gateway failed?

Every IoT dashboard should monitor:

last data timestamp,
gateway communication,
sensor availability,
and database ingestion.

A stale value should not appear as healthy current data.

33. Add Communication Health

Create panels such as:

Gateway Online
14 / 15

Sensors Online
126 / 130

Last Update
12 seconds ago

Communication Health
96.9%

This allows users to distinguish:

equipment problem

from:

communication problem

34. Use Annotations

Annotations allow events to be overlaid on time-series charts.

Useful industrial annotations include:

maintenance performed,
equipment shutdown,
sensor replacement,
electrical disturbance,
production change,
calibration,
or abnormal event.

For example, a vibration chart could display:

September 5
Bearing replaced

If vibration drops immediately afterward, the maintenance effect becomes visible.

35. Use Historical Comparison

Current trends become more useful when compared with previous behavior.

Examples include:

today vs yesterday,
this week vs last week,
current month vs previous month,
current operating cycle vs historical baseline.

This is particularly useful for:

energy monitoring,
environmental trends,
equipment condition,
and process performance.

36. Build Baselines Before Machine Learning

Industrial organizations often want predictive maintenance immediately.

But machine learning requires reliable historical data.

Before forecasting, collect:

stable measurements,
consistent sampling,
equipment operating states,
maintenance records,
fault history.

A useful maturity model is:

Dashboard

↓

Threshold Alert

↓

Trend Analysis

↓

Baseline

↓

Anomaly Detection

↓

Forecasting

↓

Predictive Maintenance

Skipping these stages usually reduces model quality.

37. Grafana Cloud Analytics

Grafana Cloud now includes several AI-assisted and analytical capabilities.

Grafana’s August 2026 overview highlights functionality including:

forecasting,
outlier detection,
Grafana Assistant,
Watchers,
and Investigations.

These can extend an industrial dashboard beyond static threshold monitoring.

38. Forecasting

Forecasting attempts to estimate future values from historical patterns.

Potential Industrial IoT applications include:

Tank Level

Forecast:

When will the tank reach the low-level limit?

Power Demand

Forecast:

Will site load exceed today’s target demand?

Energy Consumption

Forecast:

What will monthly electricity consumption likely be?

Environmental Monitoring

Forecast:

What trend is developing in monitored conditions?

Forecasting is most useful for slowly changing operational parameters.

It should not replace safety protection.

39. Outlier Detection

Outlier detection is useful where multiple similar assets exist.

Suppose ten pumps normally operate around:

3–4 mm/s vibration

but one pump operates at:

6.5 mm/s

Even if it remains below a fixed alarm threshold, the pump may be statistically unusual compared with peers.

This is a strong use case for:

pumps,
motors,
compressors,
HVAC units,
solar inverters,
environmental sensors.

40. Watchers and Automated Monitoring

Grafana’s newer monitoring workflows can further reduce the need to manually inspect dashboards continuously.

For Industrial IoT, the desired progression is:

Human watches dashboard continuously

toward:

System watches data continuously

and:

Human investigates exceptions

This can significantly improve scalability across large asset populations.

41. Use Dashboards for Root Cause Analysis

The best dashboards do more than display individual values.

They help correlate events.

Suppose a pump trips.

The engineer could examine:

13:20 Vibration increasing
13:32 Bearing temperature increasing
13:36 Motor current increasing
13:40 Flow decreasing
13:44 Pump trip

This timeline provides far more information than a final trip alarm.

Grafana becomes valuable as an engineering investigation tool.

42. Dashboard Design Principles

A technically correct dashboard can still be difficult to use.

Follow several design principles.

Put Important Information First

The top section should answer:

Is the system healthy?
What is abnormal?
What requires action?

Minimize Visual Noise

Do not display dozens of gauges simply because they look attractive.

Group Related Data

Separate:

process,
equipment health,
energy,
environment,
communication.

Use Consistent Units

Do not mix:

bar
psi
kPa

without good reason.

Use Consistent Naming

For example:

P-101 DE Bearing Vibration

rather than:

Vib1

43. Use Colors Carefully

Colors should communicate status consistently.

For example:

Normal
Warning
Alarm
No Data

The same meaning should apply throughout the monitoring platform.

Avoid dashboards where one panel uses one color to mean normal while another uses the same color to mean alarm.

44. Avoid Dashboard Overload

More data does not automatically create more insight.

An operator may need:

10 important values

rather than:

100 available sensor values

Detailed engineering data can be placed on secondary dashboards.

Use a drill-down architecture.

For example:

Site Overview

↓

System Overview

↓

Asset Dashboard

↓

Sensor Detail

This is more scalable.

45. Create Different Dashboards for Different Users

Not every user needs the same information.

Operations

Need:

current status,
alarms,
critical process parameters.

Maintenance

Need:

vibration,
temperature,
trends,
maintenance history.

Electrical Engineer

Need:

current,
voltage,
demand,
power quality.

HSE

Need:

environmental monitoring,
air quality,
weather,
exceedances.

Management

Need:

uptime,
energy intensity,
number of active problems,
site comparison,
costs.

One underlying data platform can support all these views.

46. Build Multi-Customer Architecture Carefully

For Industrial IoT service providers, customer isolation is critical.

A system may contain:

Customer A
Customer B
Customer C

Customer A should never see Customer B’s information.

Possible approaches include:

separate Grafana organizations,
separate Grafana Cloud stacks,
segregated databases,
tenant-aware queries,
controlled folders and permissions.

The right architecture depends on the scale and contractual cybersecurity requirements.

For a managed industrial monitoring service, strong tenant isolation is generally preferable to simply relying on dashboard filters.

47. Secure the Database Connection

For PostgreSQL, follow least privilege.

Grafana’s current documentation explicitly recommends a database user that has only SELECT permission on the relevant schemas and tables.

Do not configure Grafana using:

postgres superuser

or another highly privileged account.

If credentials are compromised, the impact should remain limited.

48. Use TLS

Where the database crosses networks, use encrypted communications.

Examples include:

TLS for PostgreSQL,
HTTPS for InfluxDB,
VPN,
private networking,
reverse proxy,
firewall rules.

Never expose an industrial database unnecessarily to the public internet.

49. Separate OT from Monitoring Infrastructure

An industrial architecture should avoid giving cloud dashboards direct unrestricted access to PLCs.

A safer architecture is:

PLC / Sensor Network

↓

Edge Gateway

↓

Controlled Data Interface

↓

Database

↓

Grafana

This creates several layers between external analytics and operational control.

50. Make Control and Monitoring Independent

Grafana should generally not become the primary platform for critical machine control.

Functions such as:

motor protection,
emergency shutdown,
process interlocks,
electrical protection,
and safety systems

should remain within appropriately engineered local systems.

Grafana is strongest for:

monitoring,
visualization,
analysis,
trending,
reporting,
and alerts.

A useful engineering principle is:

Control Locally. Protect Locally. Analyze Centrally.

51. Build a Dashboard for Mobile Use

Industrial users increasingly access monitoring systems from:

laptops,
tablets,
smartphones.

Dashboard layouts should therefore avoid excessive complexity.

For mobile viewing, emphasize:

equipment status,
active alerts,
critical KPIs,
essential trends.

Detailed engineering analysis is usually more effective on larger screens.

52. Optimize Database Queries

Poorly designed dashboards can generate unnecessary database load.

For example, imagine:

100 panels

each querying:

30 days of one-second data

Every refresh can become expensive.

Optimize using:

appropriate sampling,
aggregation,
downsampling,
limited time ranges,
database indexing,
retention policies.

InfluxDB’s architecture includes mechanisms such as retention and downsampling that are particularly relevant for high-volume sensor workloads.

53. Use Multiple Sampling Rates

Not all data needs to remain at one-second resolution forever.

For example:

Recent Data

1-second resolution
Retention: 7 days

Medium-Term Data

1-minute averages
Retention: 1 year

Long-Term Data

1-hour averages
Retention: several years

This dramatically reduces storage requirements.

Raw high-frequency data can be retained selectively where engineering analysis requires it.

54. Monitor the Monitoring System

A professional IoT platform must monitor itself.

Track:

gateway uptime,
database connectivity,
storage usage,
ingestion rate,
query latency,
Grafana availability,
and data freshness.

Otherwise, the monitoring system can fail silently.

A useful dashboard is:

Siteplore Platform Health

showing:

Gateways online
Database status
Grafana status
Data ingestion
API health
Storage capacity

55. Create a Standard Dashboard Template

Organizations managing many sites should standardize dashboard structure.

For example:

Page 1 — Overview

Page 2 — Equipment Health

Page 3 — Energy

Page 4 — Environmental

Page 5 — Alerts

Page 6 — Communication Health

This improves usability across installations.

Engineers moving from one customer site to another immediately understand the interface.

56. Example Complete Siteplore Dashboard

Imagine an integrated industrial facility.

The top dashboard could display:

SITE STATUS

Assets Online        118 / 120
Active Warnings      4
Critical Alarms      1
Gateways Online      12 / 12

POWERWATCH

Total Power          4.2 MW
Today's Energy       61.4 MWh
Power Factor         0.96
Peak Demand          4.8 MW

MACHINEGUARD

Healthy Assets       88
Warning              5
Critical             1

REKASENSE

Temperature          31°C
Humidity             74%
PM2.5                18 µg/m³
Rainfall             2.4 mm

COMMUNICATION

Sensors Online       278 / 282
Last Data Update     8 seconds

Users can then drill into each system.

57. Example MachineGuard Asset Detail

Select:

Asset = P-101

The dashboard could display:

Equipment

Cooling Water Pump P-101

Status

RUNNING

Health

WARNING

Measurements

Vibration DE        5.3 mm/s
Vibration NDE       4.7 mm/s
Bearing Temp DE     71°C
Motor Current       42 A
Flow                115 m³/h

Trend

30-day vibration graph.

Analysis

Vibration +28% from baseline

Action

Inspect bearing and alignment condition

That is much closer to operational intelligence than simply showing sensor readings.

58. From Dashboard to Decision Support

The evolution of an Industrial IoT dashboard should look like:

Level 1 — Visualization

Temperature = 60°C

Level 2 — Threshold

Warning above 70°C

Level 3 — Trend

Temperature increasing 1.2°C/day

Level 4 — Context

Temperature increase occurs under normal load

Level 5 — Analytics

Behavior differs from historical baseline

Level 6 — Prediction

Likely to exceed warning threshold within six days

Level 7 — Decision Support

Recommended inspection during next maintenance window

This is where Industrial IoT begins creating significant operational value.

59. Recommended Implementation Workflow

A practical project can follow this sequence.

Step 1 — Identify Use Case

Example:

Monitor cooling-water pump condition.

Step 2 — Select Parameters

vibration,
temperature,
motor current,
flow.

Step 3 — Connect Sensors

Use:

RS485,
analog inputs,
digital inputs,
or existing PLC data.

Step 4 — Configure Edge Gateway

Poll the measurements.

Step 5 — Store Data

Send data to:

InfluxDB,
PostgreSQL / TimescaleDB.

Step 6 — Connect Grafana

Configure the appropriate data source.

Step 7 — Build Dashboard

Create:

current values,
trends,
equipment status.

Step 8 — Establish Baseline

Collect normal operating data.

Step 9 — Configure Alerts

Begin with high-value conditions.

Step 10 — Add Analytics

Introduce:

trend analysis,
anomaly detection,
forecasting

only when sufficient historical data exists.

60. Common Mistakes to Avoid

Several mistakes repeatedly occur in Industrial IoT dashboard projects.

Connecting Everything Immediately

Start with critical assets.

Too Many Panels

Focus on decisions.

Too Many Alerts

Prioritize actionable events.

No Baseline

Historical context is essential.

Inconsistent Units

Standardize engineering units.

Poor Asset Naming

Establish an asset hierarchy.

Ignoring Data Quality

Monitor sensor and communication health.

Using Administrator Database Accounts

Use least privilege.

Direct Cloud-to-PLC Connectivity

Maintain architectural separation.

Treating Grafana as the Control System

Keep critical control and protection local.

Grafana, Siteplore, and Scalable Industrial Monitoring

For a platform such as Siteplore, Grafana provides an effective visualization and analytics layer above the data infrastructure.

Siteplore can connect industrial assets through:

RekaSense Edge Gateways,
RS485 Modbus RTU,
Modbus TCP,
environmental sensors,
electrical meters,
machine-condition sensors,
existing PLC data,
and industrial instruments.

The resulting data can be stored using architectures such as:

InfluxDB

or:

PostgreSQL / TimescaleDB

Grafana can then provide customer-facing dashboards for:

RekaSense

Environmental and IoT monitoring.

PowerWatch

Electrical and energy monitoring.

MachineGuard

Condition monitoring and predictive-maintenance applications.

A customer monitoring architecture can consequently evolve from:

Single Sensor

to:

Single Asset

to:

Production Area

to:

Complete Facility

to:

Multi-Site Portfolio

without changing the fundamental dashboard philosophy.

Conclusion

Building an Industrial IoT dashboard with Grafana is not primarily a visualization exercise.

It is a systems-engineering exercise.

The complete architecture must connect:

Field Measurement

↓

Communication

↓

Edge Processing

↓

Data Storage

↓

Visualization

↓

Alerts

↓

Analytics

↓

Operational Action

Grafana provides a powerful interface between industrial data and the people who need to use that data.

But dashboard success depends on several disciplines working together:

instrumentation,
electrical engineering,
industrial communication,
networking,
database design,
cybersecurity,
reliability engineering,
maintenance,
operations,
and data analytics.

Start with a clearly defined operational problem.

Build a structured asset hierarchy.

Use a database appropriate for time-series information.

Connect Grafana securely.

Create a simple overview.

Use dynamic variables to make dashboards scalable.

Visualize trends rather than only current measurements.

Add engineering context.

Configure actionable alerts.

Monitor data quality.

Then introduce anomaly detection and forecasting only after reliable historical data has been established.

When implemented systematically, Grafana can evolve from a dashboard platform into an important part of an Industrial IoT decision-support system.

The final objective is not:

More Charts.

It is:

Better Visibility.

Earlier Detection.

Better Decisions.

More Reliable Operations.

Current Grafana References

Grafana Dashboards Documentation. Grafana dashboards organize one or more panels that query, transform, and visualize information from connected data sources.

Grafana Data Sources. Grafana can connect directly to external data stores and query data where it already resides, allowing different sources to appear in a common dashboard without forcing data migration.

Grafana InfluxDB Data Source. Grafana currently supports InfluxDB OSS, Enterprise, and Cloud variants and multiple query languages depending on the InfluxDB version. The integration supports metrics, logs, alerting, annotations, transformations, template variables, and dashboards.

Grafana PostgreSQL Data Source. PostgreSQL connectivity is built into Grafana, and Grafana recommends restricting the connected database account to the minimum required SELECT permissions for production environments.

Grafana Alerting. Grafana Alerting evaluates configured data queries and expressions to detect conditions requiring attention, while contact points and notification policies control delivery of notifications.

Grafana Dashboard Variables. Variables allow one interactive dashboard to filter across environments, sites, assets, or other dimensions rather than requiring many static copies.

Grafana Cloud AI and Analytics. As of August 2026, Grafana Cloud’s AI-assisted workflows include capabilities around forecasting, outlier detection, Assistant, Watchers, and Investigations.

How to Build an Industrial IoT Dashboard with Grafana

What Is Grafana?

1. Start with the Industrial Problem

Machine Condition Monitoring

Electrical Monitoring

Environmental Monitoring

Water Monitoring

2. Define the Complete IoT Architecture

Layer 1 — Sensors and Field Devices

Layer 2 — Field Communication

Layer 3 — Edge Gateway

Layer 4 — Database

Layer 5 — Grafana

3. Choose the Right Database

InfluxDB for Industrial IoT

PostgreSQL and TimescaleDB

4. Example Siteplore Architecture

5. Design the Data Model Before Designing the Dashboard

6. Use Consistent Tagging

7. Connect Grafana to the Data Source

8. Test the Data Source

9. Build Your First Dashboard

10. Start with an Overview Dashboard

Site Overview

11. Example Dashboard Structure

Row 1 — Site Status

Row 2 — Critical KPIs

Row 3 — Trends

Row 4 — Equipment Detail

Row 5 — Analytics

12. Use the Stat Panel for Current Status

13. Use Time Series Panels for Trends

14. Combine Related Variables

15. Use Gauge Panels Carefully

16. Use Thresholds

17. Add Units

18. Create Dashboard Variables

19. Recommended Industrial Variables

20. Example Dynamic Dashboard

21. Use Repeating Panels

22. Create an Equipment Health Dashboard

Header

Current Condition

Historical Trends

Operational Context

Alerts

23. Example PowerWatch Dashboard

System Overview

Power Quality

Demand

Historical Trends

Events

24. Example RekaSense Dashboard

Weather

Air Quality

Environmental Status

Historical Trends

Location

25. Example Water Monitoring Dashboard

Hydraulic

Water Quality

Pump Condition

Energy

Alerts

26. Build Engineering Calculations into the Dashboard

27. Data Transformations

28. Create Alerts

29. Avoid Simple Alarm Logic Where Possible

30. Separate Alert Rules from Notification Routing

Electrical Alarm

Vibration Alarm

Environmental Alarm

Critical Site Alarm

31. Design Alert Severity

Information

Warning

Alarm

Critical

32. Monitor Data Freshness

33. Add Communication Health