Conditional dependency and status of step.

[bpics]

1. Is it possible to get the status of any step in the dag ? Like I want to know if step1 is successful or failure in another step execution.

2. Is it possible to choose dependency conditionally. Let's say step2 and step3 are dependent on step1. I want to execute step if step1 is successful and step3 if step1 is failure.

[yohamta]

Hello. Thanks for the question.

1. Yes via HTTP API.
2. No direct feature, but there're ways to achieve the same effect.

Example:

steps:
  - name: step1
    command: bash
    script: |
      [run some important task]
      echo "SUCCESS"
    output: STEP1_RESULT
    continueOn:
      failure: true
  - name: step2
    command: echo "step2"
    preconditions:
      - condition: "$STEP1_RESULT"
        expected: "SUCCESS"
    depends:
      - step1
  - name: step3
    command: echo "step3"
    preconditions:
      - condition: "$STEP1_RESULT"
        expected: "FAILURE"
    depends:
      - step1

[bpics]

For (1), When will the status of a step be available using HTTP call? Is it available post DAG execution or during DAG's execution? I am interested in checking the status of previous step in some step below in the execution path, within the same DAG.

For(2) - I am already executing a command, which prints n number of lines to the console. Some through print statements and some through logging mechanism. Couple of questions here:
a. I understand echo is populating the value to the variable mentioned in the output and echo also displays the text onto console. With multiple lines already present in the console how would theoutput will identify which value to pick up so that it can be used in the next steps?
b. The command has some task to do. Can the script rely on the execution status of the task mentioned by command and populate the output variable based upon that?

[yohamta]

Hi @bpics sorry for the late response.

• HTTP API returns always the latest and realtime status of the DAG (running / stopped).

• a) OUTPUT only captures standard output as is, so if you need to identify the status of the result, you may need to have an intermediate step to capture the last line of the OUTPUT. For example, you can add a step to run tail -n 1 and capture it to OUTPUT again.

• b) When a step has a script field, Dagu just runs [command] [script (temp file)]. It's nothing different from, for example running a script file named a.sh and specify bash a.sh in command field.