Wrong rendering for Rel_R

[Oleg3n]

Hi With code like this:

@startuml C4_test2
!include https://raw.githubusercontent.com/plantuml-stdlib/C4-PlantUML/master/C4_Container.puml

System_Boundary(feature2, "Feature 2") {
    Container(ms2, "ms2", "Feature")
}

Lay_D(ms1, ms2)

Rel(ms1, kafka, "Publish")
Rel_L(ms2, kafka, "Publish")

Rel_R(con1, kafka, "Consume")
Rel_R(kafka, con2, "Consume")

@enduml

I am getting the below image

Shouldn't "kafka" be on the right side from "con 1" ? Thanks in advance

[kirchsth]

Hi @Oleg3n,

did you read https://github.com/plantuml-stdlib/C4-PlantUML/blob/master/LayoutOptions.md#layout-guidance-and-practices ?

I think you have a sample which is not perfect rendered, but I think if you remove the Lay_D(ms1, ms2) it produces a better (your expected?) result.

BR Helmut

[Oleg3n]

Hi Helmut Thank you for the quick reply. Yes, I did read. Do you mean I break some rule? Or that I should have understood that drawing results are not predictable? :) Your trick is working for this minimal working test. What I initially tried to build is a diagram similar to this but vertically mirrored (DBs on the right, Externals on the left):

and the best result I was able to achieve is this:

So I am trying to understand if I am doing something wrong or is it 'expected behavior'.

Code for the first image

@startuml C4_test3
!include https://raw.githubusercontent.com/plantuml-stdlib/C4-PlantUML/master/C4_Container.puml

System_Boundary(connectors, "Connectors") {
    Container(sink1, "Sink 1", "Sink connector")
    Container(sink2, "Sink 2", "Sink connector")
}

System_Ext(ext1, "External 1", "External System")
System_Ext(ext2, "External 2", "External System")

ContainerQueue(kafka, "Kafka")

System_Boundary(feature1, "Feature 1") {
    Container(ms1, "Component 1", "Feature")
    ContainerDb(db1, "DB")
}
System_Boundary(feature2, "Feature 2") {
    Container(ms2, "Component 2", "Feature")
    ContainerDb(db2, "DB")
}
System_Boundary(feature3, "Feature 3") {
    Container(ms3, "Component 3", "Feature")
    ContainerDb(db3, "DB")
}

Lay_D(sink1, sink2)
Lay_D(ms1, ms2)
Lay_D(ms1, ms2)
Lay_D(ms2, ms3)
Lay_D(ms2, ms3)

Rel_L(sink1, ext1, "Publish")
Rel_L(sink2, ext2, "Publish")

Rel(sink1, kafka, "Consume")
Rel_R(sink2, kafka, "Consume")

Rel(ms1, kafka, "Publish")
Rel_L(ms2, kafka, "Publish")
Rel(ms3, kafka, "Publish")

Rel_L(ms1, db1, "Persist")
Rel_L(ms2, db2, "Persist")
Rel_L(ms3, db3, "Persist")

@enduml

Code for the second image. The only difference from the first is _R replaced with _L and vice versa.

@startuml C4_test1
!include https://raw.githubusercontent.com/plantuml-stdlib/C4-PlantUML/master/C4_Container.puml

System_Boundary(connectors, "Connectors") {
    Container(sink1, "Sink 1", "Sink connector")
    Container(sink2, "Sink 2", "Sink connector")
}

System_Ext(ext1, "External 1", "External System")
System_Ext(ext2, "External 2", "External System")

ContainerQueue(kafka, "Kafka")

System_Boundary(feature1, "Feature 1") {
    Container(ms1, "Component 1", "Feature")
    ContainerDb(db1, "DB")
}
System_Boundary(feature2, "Feature 2") {
    Container(ms2, "Component 2", "Feature")
    ContainerDb(db2, "DB")
}
System_Boundary(feature3, "Feature 3") {
    Container(ms3, "Component 3", "Feature")
    ContainerDb(db3, "DB")
}

Lay_D(sink1, sink2)
Lay_D(ms1, ms2)
Lay_D(ms1, ms2)
Lay_D(ms2, ms3)
Lay_D(ms2, ms3)

Rel_R(sink1, ext1, "Publish")
Rel_R(sink2, ext2, "Publish")

Rel(sink1, kafka, "Consume")
Rel_L(sink2, kafka, "Consume")

Rel(ms1, kafka, "Publish")
Rel_R(ms2, kafka, "Publish")
Rel(ms3, kafka, "Publish")

Rel_R(ms1, db1, "Persist")
Rel_R(ms2, db2, "Persist")
Rel_R(ms3, db3, "Persist")

@enduml

[kirchsth]

Hi @Oleg3n,

I fear it is basically the "expected" behavior and I have no better rules.

I tried it with a different LAYOUT_LANDSCAPE(), changed the order of definitions, moved internal relations into the system boundaries, removed some double defined LAY_D() calls (and used feature1 instead of ms1, because I think the layout of the outside boundaries could be better(?)) and got something like this.

BR Helmut

[Oleg3n]

Thank you, Helmut. This is really helpful. This unpredictability for _L and _R makes me crazy but this is good news that eventually it should be possible to achieve the desired result. I appreciate your help and effort. All the very best in this New year Oleg