Control and monitoring items

An item in EVA means any object which can be controlled or monitored. Universal Controller has 2 native item types: unit and sensor, plus multi-updates, which can contain multiple items at once.

Logic Manager has one native item type lvar (logic variable), additionally it loads remote units and sensors from connected UCs. SCADA Final Aggregator has no native item types and loads everything from the connected remote controllers.

The configurations of items are stored in runtime folder, from where the controllers loads .json files. Each item has multiple parameters which can be predefined or customized. All customized parameters can be displayed with the help of API list_props or EVA console tools In case the configuration file settings are changed manually or by 3rd party software, controller should be restarted to reload the configurations.

Common item parameters

  • id item ID, e.g. ‘lamp1’. When using simple layout, must be unique within one controller, even if items are in different groups. This creates some complications when designing the whole installation architecture but allows to keep EVA configuration and item scripts organized in a simple way and makes system administration and support much easier.

  • group item group, i.e. ‘hall/lamps’. Assigned at the time of item creation, the group can’t be changed later to avoid synchronization problems.

  • full_id full item id (i.e. ‘hall/lamps/lamp1’), read-only. Must be unique within one controller despite of layout used.

  • oid object id, unique within the whole installation, same as full_id, but also contains the item type: ‘unit:hall/lamps/lamp1’, read-only

  • description item description


All EVA functions, commands and parameters can accept oid as the item identifier.

Item layout

EVA ICS v3 allows to use two types of item layouts: simple and enterprise. if the system is not being used yet and there are no items created, layout can be set in Universal Controller and Logic Manager configuration files (option within [server]: layout=<simple|enterprise>). Otherwise you should use sbin/layout-converter tool to convert simple layout to enterprise. Inverse conversion is not possible.

Benefits of simple layout:

  • Good to use in simple installations or in the installations where each component has no similar items. Each item should have its own unique ID, despite that items are located in different groups.

  • When doing controller maintenance tasks, you can address each item by its ID instead of full ID or oid.

  • Item configuration files are named as <ID>.json and can be easily located.

Benefit of enterprise layout: different items in different groups can have the same IDs. Ideal for setups where multiple similar components are managed by one controller.

In general, simple layout should be used only for testing and simple temporary setups. For usage in production environment, enterprise layout is always recommended.


A unit is a physical item, a device that we control. A unit is not a relay port, a dimmer or a controlled resistor. This is an object, for example: an electric lamp chain, a door, ventilation, a window, a pump or a boiler.

The unit can be controlled with one relay (e.g. a lamp chain: we control the whole chain by turning on/off the relay port) or with several ones (controlling e.g. a garage door often requires two relays: the first one starts the motor, the second one chooses the direction of movement). However, a door is one unit with “open” or “closed” statuses.

All units are connected to Universal Controller subsystems, which control them and form the single “unit” with one or several relays/programmable switches using control scripts. One Universal Controller can work with multiple units, but one unit should be connected to only one Universal Controller in order to avoid conflicts. Nevertheless, for reliability, one unit can be connected to several controllers, if its state is correctly synchronized via MQTT.

Each unit has its unique ID, for example “lamp1”. ID can include numbers, uppercase and lowercase Latin characters and some special characters like minus (-) or dot (.).

Unit parameters are set via configuration.

Status of the unit state

Status of the unit state is always an integer (a positive number or 0), and is by default 0 - unit is “off” (inactive) and 1 - “on” (active).

A unit can have other statuses: for example, a dimmer can include status 2 - enabled at 10% of the capacity, 3 - enabled at 50% of the capacity, window can be fully open or 50%. In the item configuration, you can assign a label to each status for enhancing its usability in interfaces.

Status -1 indicates that unit has an error status. It is set from the outside or by the system itself if the unit wasn’t updated for more than “expires” (value from item config) seconds.

Value of the unit state

Sometimes it’s not necessary to create multiple new statuses for the unit. In such cases, the unit also has a “value” parameter (which can include both numbers and letters). For instance, a motor can be controlled by two unit statuses - 0 and 1, i.e. turned on/off, but Its speed is set by value. You can also use value to control e.g. dimmers.

EVA does not use unit value for internal control and monitoring logic (except in your custom macros), that is why you can set it to any value or several values separating them with special characters for further processing.

Units in EVA hive

All units have OIDs like unit:group/unit_id e.g. unit:light/room1/lamp1

For synchronization via MQTT, the following subjects are used for units

  • [space/]unit/<group>/<unit_id>/status unit status, integer

  • [space/]unit/<group>/<unit_id>/value unit value

  • [space/]unit/<group>/<unit_id>/nstatus next unit status (different from status if action is being executed), integer

  • [space/]unit/<group>/<unit_id>/nvalue next unit value

  • [space/]unit/<group>/<unit_id>/action_enabled are actions enabled for the unit or not (boolean, True/False)

Unit parameters

  • expires integer value, time (seconds) after which the item state is considered “expired”. If the item state was not updated during this period, the state automatically is set to -1 (error), value is deleted (set to null). If ‘expires’ param is set to 0, this feature is disabled. The minimum expiration step is 0.1 sec.

  • mqtt_update = “notifier:qos” if set, the item can receive active state updates through notification from the specified MQTT server. Example: “eva_1:2”. State updates should be sent either to MQTT topics “path/to/unit/status” and “path/to/unit/value” or as JSON message to “path/to/unit”. In example, to set unit “tests/unit1” status to 1:

    • MQTT topic: unit/tests/unit1

    • MQTT payload:

      { "status": 1 }

    There is also a configuration field server/mqtt-update-default which can be set in config/uc/main registry key (default e.g. to eva_1:2) and applied to all newly created items.

  • snmp_trap if set, the item can receive active state updates via Connecting equipment with drivers.

  • update_delay delay in seconds before triggered update is launched

  • update_exec a script for passive update of the item state, “xc/uc/ITEMID_update” by default.

  • update_interval integer value, time (seconds) interval between the calls for passive update of the item. Set 0 to disable passive updates. Minimum step is 0.1 sec.

  • update_timeout integer, value, time (seconds) in which the script of the passive update should finish its work or it will be terminated.

  • action_allow_termination boolean, allow currect running action termination by external request.

  • action_always_exec boolean, always execute the actions, even if the intended status is similar to the current one

  • action_enabled boolean, allow or deny new actions queue/execution

  • action_exec a script which performs the action, “xc/uc/ITEMID” by default.

  • action_queue={0|1|2}

    • 0 action queue is disabled, if the action is running, new actions are not accepted

    • 1 action queue is enabled, all new actions are put in queue and executed in a normal way

    • 2 queue is disabled, new action terminates the current running one and then is executed

  • action_timeout integer, value, time (seconds) in which the script of the action should finish its work or it will be terminated.

  • auto_off integer, the simple automation parameter: the command to turn the unit off (call an action to set status = 0) will be executed after the indicated period of time (in seconds) after the last action performed for this unit. Set 0 to disable this feature. Minimum step is 0.1 sec.

  • location you can specify units’ physical location, as GPS coordinates or in custom format. To specify GPS coordinates, set the parameter to value longitude:latitude or longitude:latitude:altitude. If you choose to set location as GPS or some other coords, full unit state is appended with virtual parameters loc_x, loc_y (and if altitude is specified - loc_z). These virtual parameters are parsed automatically from location and can be used later e.g. to filter units by location or to put units on geographical map.

  • maintenance_duration integer, if greater than zero, item can enter maintenance mode on the specified amount of seconds. During maintenance mode all item updates are ignored, however actions still can be executed. If expires property is also set, item state expire after maintenance_end + expires seconds.

  • mqtt_control = “notifier:qos” item gets actions through notifications from a specified MQTT server, for example “eva_1:2”, actions should be sent to path/to/unit/control (e.g. unit/hall/lamps/lamp1/control).

    The message should be:

    • either in a form of text messages “status [value] [priority]”. If you want to skip value, but keep priority, set it to null, i.e. “status 0 null 50”.

    • or in JSON format (fields “value” and “priority” are optional):

      { "status": 1, "value": "", "priority": 100 }
  • modbus_status, modbus_value update item state from Modbus slave memory space.

  • notify_events 2 (default) - send notifications about all events, 1 - about state only, 0 - disable all event notifications.

  • status_labels “labels” used to display the unit statuses by the interfaces. Labels can be changed via UC API or eva uc, in the following way: status:number = label, e.g. “status:0” = “stop”. By default the unit has labels “status:0” = “OFF”, “status:1” = “ON”. Status labels can be used as status param to execute unit actions, in this case controllers check the status match to the specified label (case insensitive).

  • term_kill_interval integer, difference (in seconds) between stopping and forcible stopping the action or update script. Tip: sometimes it is useful to catch SIGTERM in the script to exit it gracefully. Cannot exceed the value of timeout** 2, where timeout** default timeout, set in a controller config.

  • update_exec_after_action boolean, start passive update immediately after the action is completed (to ensure the unit state has been changed correctly)

  • update_if_action boolean, allow or deny passive updates while the action is being executed

  • update_state_after_action boolean, if action is completed successfully, the controller assumes that its actual unit state has been changed correctly and sets it without calling/waiting for the state update.


The sensor value is the parameter measured by the sensor: temperature, humidity, pressure etc.

In terms of automation the difference between sensor item and unit item is obvious: we change the unit state by ourselves and monitor it only for the sake of checking the control operations, while the sensor state is changed by the environment.

Regarding the system itself, unit and sensor are similar items: both have status and value, the item status is monitored actively (by UC API, MQTT message, SNMP traps) or passively (by calling the external script).

The sensor can have 3 statuses:

  • 1 sensor is working and collecting data

  • 0 sensor is disabled, the value updates are ignored (this status can be set via API or by the user)

  • -1 sensor error (“expires” timer went off, the status was set because the connection with a physical sensor got lost during passive or active update etc), when the sensor is in this status, its value is not sent via notification system to let other components work with the last valid data.


The sensor error state is automatically cleared if new value data arrives.

Important: the sensor error can be set even if the sensor is disabled. It means that the disabled sensor can be switched to “error” and then to “work” mode by the system itself. Why it works that way? According to the logic of the system, the sensor error is an emergency situation that should affect its status even if it is disabled and requires an immediate attention of the user.

Sensors (and sometimes units) can be placed on the same detector, controller or bus queried by a single command. EVA can use multi-updates in order to update several items at once.

Since the system does not control, but only monitors the sensor, it can be easily connected to several Universal Controllers at once if the equipment allows making parallel queries of the state or sending active updates to several addresses at once.


The sensor doesn’t set its status to ‘-1’ on expires if its status is 0 (disabled)

Sensors in EVA hive

All sensors have OIDs like sensor:group/sensor_id e.g. sensor:temp/t1

For synchronization via MQTT, the following subjects are used for units

  • [space/]sensor/<group>/<sensor_id>/status sensor status, integer

  • [space/]sensor/<group>/<sensor_id>/value sensor value

Sensor parameters

Sensors have the same parameters as units, except they don’t have action_*, auto_off, mqtt_control, modbus_status and status_labels.

Validation of state value

State value of units and sensors can be validated before Universal Controller perform item update.

To validate item value, the following item properties are used:

  • value_in_range_max value should be less

  • value_in_range_max_eq value should be less or equal than specified max.

  • value_in_range_min value should be greater

  • value_in_range_min_eq value should be greater or equal than specified max.

or virtual parameter value_condition, which can be set in human readable way, e.g. 20<x<=200.

If value validation is set and item receive state value which is not numeric or doesn’t feet the specified range, item keep previous state value and get status -1 (error).

Item status is set back to normal as soon as any valid state update is received.

Logic variable

EVA Logic Manager uses the logic variables (lvars) to make decisions and organize production cycle timers.

The parameters of logic variables are set in their configurations.

Actually lvars are similar to sensors, but with the following differences:

  • The system architecture implies that the sensor value is changed depending on the environment; the logic variables are set by the user or the system itself.

  • The logic variables, as well as the sensors, have statuses -1, 0 and 1. However, if the status is 0 (variable is disabled) it stops responding to any value-only changes.

  • The logic variables exchange two more parameters with the notification system: “expires” (time in seconds after the variable is set, and then takes the null value and -1 status) and set_time - time when the value was set for the last time.

The same logic variable can be declared on several logic controllers, but the “expires” configuration value should remain the same because each controller processes it autonomously. The variable becomes “expired” once it is declared as such by any controller.


LVar doesn’t set its status to ‘-1’ on expires if its status is 0 (disabled)

The logic variable values can be synchronized via MQTT server or set via API or external scripts - similar to sensors.

You can use several logic variables as timers in order to organize production cycles. For example, there are three cycles: the pump No.1 operates in the first one, the pump No. 2 in the second one, and both pumps are disabled in the third one. In order to organize such cycle, let us create three variables: cycle1, cycle2, cycle_stop with “expires” values equal to the duration of each cycle in seconds.

Then - in the decision-making matrix you should specify the rules and macros run as soon as each cycle is finished. The macros run and stop the pumps as well as reset the timer variables of the next cycle: as soon as cycle_stop is finished, the pump No.1 is run, the cycle1 timer variable is reset; as soon as the cycle1 is finished, the pump No. 2 is run and cycle2 variable is reset; as soon as cycle2 is finished, both pumps are disabled and cycle_stop is reset.

In order to synchronize timer values with interfaces and the third-party applications, use LM API test command that displays the system information, including local time on the server on which the controller is installed.

However, when used in industrial configurations, it is recommended to synchronize time on all computers without any additional software hotfixes.

LVars in EVA hive

All logic variables have OIDs like lvar:group/lvar_id e.g. lvar:service/var1

For synchronization via MQTT, the following subjects are used for units

  • [space/]lvar/<group>/<lvar_id>/status lvar status, integer

  • [space/]lvar/<group>/<lvar_id>/value lvar value

  • [space/]lvar/<group>/<lvar_id>/set_time last set time (Unix timestamp)

  • [space/]lvar/<group>/<lvar_id>/expires value expiration time (seconds)

Examples using LVars

You can use lvar as a

  • Variable To use lvar as a shared variable to exchange some information between controllers, apps and SCADA interfaces, just set its value (and status if you want) and that’s it.

  • Timer

    • Set expires configuration param

    • Use reset to set lvar status/value to 1 and reset the expiration timer

    • Use clear to set lvar status to 0 and stop it reacting to expiration (when used with lvar which have expires param set, clear changes its status instead of value)

    • Use decision rules with the conditions on_set and on_expire to run the macros when the timer is set/expired

    • if the timer has status set to 1, it’s running

    • if status is 0, it’s disabled with clear function

    • if status is -1 and value is null (empty), the timer is expired

  • Flag

    • Use lvar as a simple boolean variable to exchange the information True/False, yes/no, enabled/disabled etc.

    • Use reset to set lvar value to 1 which should be considered as True

    • Use clear to set lvar value to 0 which should be considered as False

    • Use toggle to toggle lvar value between 0 and 1

    • Use constructions like if value(‘lvar_id’): in macros to determine is the ‘flag’ lvar is set or not.

LVar parameters

As LVars behavior is similar to sensors except the values are set by user/system, they have the same parameters, except lvars can’t be updated via SNMP traps / MQTT.

LVars also have an additional parameter logic.

Simple logic

If lvar property logic is set to ‘simple’, it bypasses built-in logic behavior conditions:

  • LVar status can be set to any integer

  • Status “0” is considered as normal and:

    • Lvar value can be updated even if LVar status is 0

    • LVar is considered as expired even if LVar status is 0

  • When LVar becomes expired, its value is cleared, but the status is kept

Simple logic is useful to use LVars as simple double data holding registers (integer + any value) for various tasks.



Starting from EVA ICS 3.3.2, multi-updates are deprecated

Multi-updates allow Universal Controller updating the state of several items with the use of one script. This could be reasonable in case all items are placed on the same bus or external controller and queried by a single command.

Multi-update is an independent item in the system with its own configuration and without status and value. In turn, it updates statuses of the included items.

Multi-updates in EVA hive

All multi-updates have OIDs like mu:group/mu_id e.g. mu:environment/mu1

Multi-updates don’t have their own state, so they are not synchronized between servers.

Multiupdate parameters

Multi-updates have the same parameters as sensors, except that “expires”, “mqtt_update” and “snmp_trap”, plus some additional:

  • items = item1, item2, item3… - the list of items for updating, can be changed via UC API and eva uc as follows:

    • -p “item+” -v “item_id” add item for update

    • -p “item-” -v “item_id” delete item

    • -p “items” -v “item1,item2,item3…” replace the whole list

  • update_allow_check - boolean, the multi-update will be performed only in case the passive state updates are currently allowed for all included items (i.e. if some of them run actions at this moment and have update_if_action=False, multi-update will be not executed)


Multiple CVARs, units and sensors can be merged in logical groups called devices. It’s completely up to you how to merge items into device, but it’s recommended to keep them in one or several separate item groups.

Device templates are stored in runtime/tpl folder in YAML (default) or JSON format.

You can use uc-tpl command line tool to create device templates using the existing items and eva uc or device management UC API functions to create, update and destroy devices.

Device management requires master key or a key with allow=device permission.


Starting from EVA ICS 3.3.2, device template format is equal to IaC and deployment files. The old format is deprecated.

Device example

Let’s imagine we have some hardware device, which has 1 relay and 2 sensors. We have a lot of devices like this and we want to create them using template.

Create one instance of device in Universal Controller defining all its items:

  • sensor:device1/device1.sensor1

  • sensor:device1/device1.sensor2

  • unit:device1/device1.relay1

Configure all defined items, then run:

uc-tpl generate -i sensor:device1/device1.sensor1,sensor:device1/device1.sensor2,unit:device1/device1.relay1

This will output device JSON template. Use -t param to output template to file or copy/paste it from the screen. You can use -c param to ask the tool automatically prepare template variables, but in our example it will just replace all 1 to {{ ID }}. We don’t want it to be done this way because we have sensor1 and relay1 items, so let’s edit the template manually:

    "sensors": [
            "group": "device{{ ID }}",
            "id": "device{{ ID }}.sensor1"
            "group": "device{{ ID }}",
            "id": "device{{ ID }}.sensor2"
    "units": [
            "group": "device{{ ID }}",
            "id": "device{{ ID }}.relay1"

(template will also contain items’ configurations which are omitted in the example)

Save the final template as runtime/tpl/mydevice.json folder, and then

# execute this command to create new device "device5"
eva uc device create mydevice -C ID=5 -y
# execute this command to destroy "device5"
eva uc device destroy mydevice -C ID=5

Configurations of the newly created items of device5 are exact copies of the items of device1. The only configuration difference is the params where we’ve put template variables instead of part or full param value (in our example: {{ ID }}).


Device templates are actually jinja2 templates, so you can use any jinja2 syntax in them (loops, conditions and etc.)

Device limitations


It is possible to re-define default initial item configuration by using “config/uc/defaults” (for units and sensors) and “config/lm/defaults” (for lvars) registry keys.

E.g. let us make units to have actions enabled by default, when created and sensors, enabled by default (when created or when there is no state in the state database).

eva-registry edit eva3/$(hostname)/config/uc/defaults
    action_enabled: true
    status: 1
eva uc server restart
eva uc create unit:tests/unit1


Changing defaults is considered as an advanced operation, if defaults are set to incorrect values, controllers do not create items properly.