Streaming Analytics

Release 10.18.0.10

This release upgrades the Apama-ctrl microservice to use Apama 10.15.4.2 (which is the same as Apama 10.15 Fix 18).

Fixes

Release 10.18.0.9

In this release, the Apama-ctrl microservice uses the same Apama version as in the previous 10.18.0.8 release.

Fixes

Release 10.18.0.8

In this release, the Apama-ctrl microservice uses the same Apama version as in the previous 10.18.0.7 release.

Fixes

Release 10.18.0.7

In this release, the Apama-ctrl microservice uses the same Apama version as in the previous 10.18.0.6 release.

Fixes

Release 10.18.0.6

In this release, the Apama-ctrl microservice uses the same Apama version as in the previous 10.18.0.5 release.

Fixes

Release 10.18.0.5

This release upgrades the Apama-ctrl microservice to use Apama 10.15.4.1 (which is the same as Apama 10.15 Fix 17).

Fixes

Release 10.18.0.4

This release upgrades the Apama-ctrl microservice to use Apama 10.15.3.3 (which is the same as Apama 10.15 Fix 14).

Fixes

Release 10.18.0.3

In this release, the Apama-ctrl microservice uses the same Apama version as in the previous 10.18.0.2 release.

Fixes

Release 10.18.0.2

This release upgrades the Apama-ctrl microservice to use Apama 10.15.3.1 (which is the same as Apama 10.15 Fix 12).

Fixes

Release 10.18.0

Apama correlator version

This release of Cumulocity IoT Streaming Analytics includes the Apama version 10.15.3 correlator. EPL apps developers should also refer to What’s New In Apama 10.15.3 in the Apama documentation.

Microservice SDK and Web SDK

Streaming Analytics has been upgraded to use build version 10.18.0.60 of the Microservice SDK and Web SDK.

Restructured Streaming Analytics guide

The Analytics Builder documentation, which is available as a separate webhelp until release 10.16, has been moved to the Streaming Analytics guide. The Streaming Analytics guide has been restructured to reflect this change. Analytics Builder developers can now search and reference the Analytics Builder documentation within the Cumulocity IoT documentation. This makes it easier for new users to get started and for advanced users to use a single documentation interface.

Note that the path to the Streaming Analytics guide has changed. The apama folder has been removed and the guide is now located in the new streaming-analytics folder. If you have bookmarks to Streaming Analytics guide, you have to update them.

As of Cumulocity IoT release 10.17, the Analytics Builder documentation is no longer provided as a separate webhelp.

New multi-tenancy microservice

The new Apama-ctrl-mt-4c-16g microservice supports multi-tenancy, but has limitations in this version. Currently, it only supports EPL apps and smart rules and does not support Analytics Builder. Furthermore, EPL apps are only enabled on the Management and Enterprise tenants that own the microservice, but disabled on the subtenants. See also Microservice runtime and applications in the Streaming Analytics guide.

Improvements in alarms generated by the Apama-ctrl microservice

The following improvements are available in this release:

• input_queues_filling, output_queues_filling, and input_output_queues_filling
  These are new types of performance alarms. They distinguish between problems with the different types of queues and replace the application_queue_full alarm type, which has been removed in this release. See Performance alarms in the Streaming Analytics guide for detailed information on the new alarms.

• parent_tenant_not_subscribed
  This is a new alarm type which is raised for a subtenant that was subscribed before the parent tenant was subscribed. See Parent tenant not subscribed in the Streaming Analytics guide for detailed information on this new alarm.

New EPL library

A new EPL library has been added to support programming in a functional style for quicker development, fewer for loops, and more concise and readable code. For example:

// Calculate the average of the values in the sequence = 20
Fn.reduce([10, 20, 30], Fn.mean());

// Filter strings by regular expression using the string.matches action = ["Hello Alice", "Hello Al"]
Fn.filter(["Hello Alice", "Hello Bob", "Hello Al"], Fn.callAction("matches", ["Hello Al.*"]));
// Use Functional() to chain multiple operations: filter out odd numbers, then reduce the sequence to a single value containing the sum = 70
Functional([10, -5, 20, 40]).filter(Fn.even).reduce(Fn.sum);

The library provides similar capabilities to Python’s functools module. It can be added to Apama projects with a new Functional EPL Library bundle which provides two main EPL types: com.apama.functional.Fn and com.apama.functional.Functional. These provide functional operations such as map, filter and reduce in both a functional style and a fluent style. They operate on EPL sequence and dictionary containers and on a new “generator” concept that lazily calculates infinite lists.

There is support for partial binding of action arguments:

Fn.map(["Bob", "Alice"], Fn.partial(Fn.concat, "Hello ")).toString(); // = ["Hello Bob","Hello Alice"]

There are also some new actions to make event sending and listening more concise, including the following:

// Initialize an event (providing only the subset of fields you care about), send and extract the request id for use in listener
// (it is also possible to initialize event fields by name rather than by position)
on MyResponse(reqId=<integer> Fn.sendToChannel(MyEvent.SEND_CHANNEL,
        Fn.setFields(new MyEvent, [<any>createRequestId(), "/myrequest"])
        ).getEntry("reqId") ) as response
{
  ...
}

// Listen for events with an identifier matching each value from this sequence, and calls a completion action when all have arrived
// (similar to an EPL "and" event expression). A different callback is executed if the specified timeout expires.
Functional(sequenceIDs)
                 .waitForAllCompleted("MyResponse", "id", onCompleted)
                 .onTimeout(TIMEOUTSECS, onTimeout);

For detailed information, see Using Functional Operations in EPL in the Apama documentation. See also the API Reference for EPL (ApamaDoc) for detailed information on com.apama.functional.

EPL improvement for the time zone

The at() event operator in EPL for listeners can now be provided with a time zone as an optional final argument. This can be in string form or (recommended) one of the new constants in the TimeZone namespace. For more details, see Triggering event listeners at specific times in the Apama documentation and the TimeZone namespace in the API Reference for EPL (ApamaDoc). The new constants can also be used with the TimeFormat event library.

New EPL sample

A new EPL sample named “Create an alarm if a measurement exceeds a threshold value (multi-tenant)” can now be accessed from the EPL editor of the Streaming Analytics application. It shows how to write an EPL app which can be executed in a multi-tenant microservice on the parent tenant. See also Developing apps with the Streaming Analytics application in the Streaming Analytics guide.

For more detailed information, see Working with multi-tenant deployments in the Apama documentation.

Improvements in Analytics Builder

The Cron Timer block now has a Time Zone parameter to customize which time zone this block is relative to.

The Alarm Output block no longer reactivates “Acknowledged” alarms with every modification. By default, acknowledged alarms are now only reactivated if the severity has increased. This also ensures that the alarm count is updated correctly. To always activate alarms triggered by the model, select the new Always Activate Alarm parameter in the Alarm Output block.

Synchronous output blocks are now allowed to have more than one output block which generates events with any given key (the key is made up of multiple block parameters). This applies to the following blocks: Alarm Output, Event Output, and Measurement Output. See also Analytics Builder > Wires and blocks > Keys for identifying a series of events in the Streaming Analytics guide.

In the Choose Device, Group or Asset dialog box, you can now choose whether you want to receive input data from the child devices of an asset or from the asset itself. Note that a model cannot receive input from both an asset and an asset’s child devices; the model must receive from either an asset itself or its child devices. If you have custom input blocks, they may need to be updated to support receiving from the assets themselves, instead of from their child devices. For detailed information, see Update Cumulocity IoT input blocks to receive from assets in the Analytics Builder Block SDK documentation.

For the output blocks, it is now possible to set a template parameter to a trigger device. For this purpose, the type that you can select in the Template Parameters dialog box has been renamed to Device, Group, Asset or Trigger Device. Note that in the instance editor it is invalid to use a template parameter that has been defined for an input block as a trigger device. When a model uses a trigger device as an input, a runtime error will occur indicating that the validation has failed.

In previous versions, it was not possible to select a trigger device as the value of a template parameter, and you had to select a device group instead. Starting from this version, it is possible to select a trigger device as the value of a template parameter. Selecting a trigger device or a device group has the same effect in this version. However, selecting a device group will be deprecated in a future version. So if you have template parameters that currently use a device group and that are used as the output destination of an output block, you should change them now to use the trigger device.

In previous versions, the palette of the model editor showed the names of the devices and device groups that are registered in the Cumulocity IoT inventory and you had to select the block type in the block parameter editor. As of this version, these names are no longer shown in the palette. Instead, the palette now shows the input and output blocks by type. For example, the Input category now shows an input block named Alarm Input and the Output category now shows an output block named Event Output. You drag an input or output block onto the canvas and then select the input source or output destination from the block parameter editor. It is no longer possible to change the block type in the block parameter editor.

For the input blocks, it is now possible to select a smart group as the input source. See also Editing the parameters of a block in the Streaming Analytics guide. A smart group is a special group in Cumulocity IoT which is dynamically constructed based on filtering criteria. For more information, see Using smart groups in the User guide.

For the input and output blocks, it is now possible to select an asset as the input source or output destination. See also Editing the parameters of a block in the Streaming Analytics guide. Assets represent business objects in general like buildings, machines, production units or cars. For more information, see Managing assets in the User guide.

The following blocks now support multi-line input for certain parameters, which means that your input is now automatically wrapped in the text box and that you can now also press Enter to start text on a new line:

• The Alarm Output block in the Message parameter.
• The Event Output block in the Message parameter.
• The Operation Output block in the Description parameter.
• The Send Email block in the Text parameter.
• The Send SMS block in the Text parameter.
• The Text Substitution block in the Text Template parameter.
• The Constant Value block in the Value parameter.
• The Selector block in all of the Input n and No Input parameters.

This is also reflected in the Template Parameters dialog box, where you can now select Multi-line String as the type for the parameters listed above. See also Managing template parameters in the Streaming Analytics guide.

When you move the mouse pointer over the input or output port of a block, the description of this port is now displayed in the tooltip in addition to the port name.

The Latest Updated mode of the Combiner block, which has been added in release 10.16.0, has been renamed to Latest Input. This was done so that you can better distinguish it from the Latest Changed mode.

First-time activation of Analytics Builder

In some circumstances, Analytics Builder must be activated by the user before it is used for the first time. If you see the Activate button for Analytics Builder on the home screen of the Streaming Analytics application, clicking this will start the process. This process may take a few minutes.

Fixes