jenkinsci / jenkinsci/blueocean-plugin
[JENKINS-48859] BlueOcean API parallel node duration in ms incorrect in declarative pipeline
- Dominant language
- Java
- Stars
- 2.9k
- Forks
- 533
- PR merge metrics
- No merged PRs in 30d
Description
When moving our pipeline from scripted to declarative, we've noticed that the node duration in milliseconds returned by the blue ocean rest API appears to be incorrect for parallel stages. It seems it may be reporting just the time to start the stage rather than the total duration.
Example pipeline:
pipeline {
agent any
stages {
stage('parallel sleep') {
parallel {
stage('1') {
steps {
sleep time: 1, unit: 'MINUTES'
}
}
stage('2') {
steps {
sleep time: 1, unit: 'MINUTES'
}
}
}
}
}
}
Blue Ocean API (/blue/rest/blueocean-api-bug-test/runs/2/nodes/) returns:
"displayName": "1",
"durationInMillis": 171,
"id": "9",
"input": null,
"result": "SUCCESS",
"startTime": "2018-01-09T09:13:18.778+0000",
"state": "FINISHED",
Compared to an example in scripted:
node('m1.medium') {
stage('parallel sleep') {
parallel([
'1': {
sleep time: 1, unit: 'MINUTES'
},
'2': {
sleep time: 1, unit: 'MINUTES'
},
])
}
}
Blue Ocean API (/blue/rest/blueocean-api-bug-test/runs/3/nodes/) returns:
"displayName": "1",
"durationInMillis": 60274,
"id": "9",
"input": null,
"result": "SUCCESS",
"startTime": "2018-01-09T09:53:00.085+0000",
"state": "FINISHED",
It feels like a bug to me. If not a bug, then it feels like an inconsistency in behaviour unless I have done something horribly wrong in my implementation (likely! 
)
---
Originally reported by rjbwtr, imported from: BlueOcean API parallel node duration in ms incorrect in declarative pipeline
Raw content of original issue
When moving our pipeline from scripted to declarative, we've noticed that the node duration in milliseconds returned by the blue ocean rest API appears to be incorrect for parallel stages. It seems it may be reporting just the time to start the stage rather than the total duration.
Example pipeline:
pipeline {
agent any
stages {
stage('parallel sleep') {
parallel {
stage('1') {
steps {
sleep time: 1, unit: 'MINUTES'
}
}
stage('2') {
steps {
sleep time: 1, unit: 'MINUTES'
}
}
}
}
}
}Blue Ocean API (/blue/rest/blueocean-api-bug-test/runs/2/nodes/) returns:
<snip>
"displayName": "1",
"durationInMillis": 171,
"id": "9",
"input": null,
"result": "SUCCESS",
"startTime": "2018-01-09T09:13:18.778+0000",
"state": "FINISHED",
<snip>Compared to an example in scripted:
node('m1.medium') {
stage('parallel sleep') {
parallel([
'1': {
sleep time: 1, unit: 'MINUTES'
},
'2': {
sleep time: 1, unit: 'MINUTES'
},
])
}
}Blue Ocean API (/blue/rest/blueocean-api-bug-test/runs/3/nodes/) returns:
<snip>
"displayName": "1",
"durationInMillis": 60274,
"id": "9",
"input": null,
"result": "SUCCESS",
"startTime": "2018-01-09T09:53:00.085+0000",
"state": "FINISHED",
<snip>It feels like a bug to me. If not a bug, then it feels like an inconsistency in behaviour unless I have done something horribly wrong in my implementation (likely!
environment
```
Jenkins: CloudBees Jenkins Enterprise 2.46.24.0.2-fixed
Pipeline - Declarative: 1.2.4
Blue Ocean: 1.3.3
```
Contributor guide
Assessment
This issue has not been assessed yet.