-
Task
-
Resolution: Duplicate
-
Critical
-
-
tasman
In Scope
- Update the loading of the artifacts tab so that it fetches its data from the new artifacts endpoint built in
JENKINS-39761(similar to how we load the test tab) - Ensure that the name of the artifact is the "full path"
- Add a download all button to the Artifacts tab (see screenshot of https://ci.jenkins.io/job/Plugins/job/pipeline-model-definition-plugin/job/master/lastSuccessfulBuild/artifact/)
- If we are showing the maximum number of artifacts then show message below "Showing first 100 artifacts"
- duplicates
-
-
- Resolved
-
- is blocked by
-
-
- Resolved
-
- is blocking
-
-
- Closed
-
- relates to
-
-
- Resolved
-
[JENKINS-39737] Update UI to support new artifacts endpoint
| Description |
Original:
When people archive large numbers of files, blue ocean shows this up as lots of items of json test, vs a heirachy that can be browsed. For example, https://ci.jenkins.io/blue/rest/organizations/jenkins/pipelines/Plugins/pipeline-model-definition-plugin/activities/?start=0&limit=1 Will return 23000 lines of json, much of which is in the artifacts field (nearly 2000 objects, each with 5 or so items). This is as a directory structure has been archived with lots of small files, but blue ocean shows this up as a flat listing. This can not scale and seems to be wrong by design. The classic jenkins provides a browsable view of artifacts: eg !Screen Shot 2016-11-15 at 7.44.41 pm.png|thumbnail! This doesn't fetch ANY artifact data until it is asked for (vs blue ocean that gives you a flat list). The raw json: https://gist.github.com/michaelneale/988cfc3ce44458ec9b094ef5dd97c9b3 The job: https://ci.jenkins.io/job/Plugins/job/pipeline-model-definition-plugin/ The pretty printed json: !Screen Shot 2016-11-15 at 7.46.33 pm.png|thumbnail! Blue ocean seems a bit wrong by design here - the artifacts tab can't be a flat listing. |
New:
When people archive large numbers of files, blue ocean shows this up as lots of items of json test, vs a heirachy that can be browsed. For example, https://ci.jenkins.io/blue/rest/organizations/jenkins/pipelines/Plugins/pipeline-model-definition-plugin/activities/?start=0&limit=1 Will return 23000 lines of json, much of which is in the artifacts field (nearly 2000 objects, each with 5 or so items). This is as a directory structure has been archived with lots of small files, but blue ocean shows this up as a flat listing. This can not scale and seems to be wrong by design. The classic jenkins provides a browsable view of artifacts: eg !Screen Shot 2016-11-15 at 7.44.41 pm.png|thumbnail! This doesn't fetch ANY artifact data until it is asked for (vs blue ocean that gives you a flat list). The raw json: https://gist.github.com/michaelneale/988cfc3ce44458ec9b094ef5dd97c9b3 The job: https://ci.jenkins.io/job/Plugins/job/pipeline-model-definition-plugin/ The pretty printed json: !Screen Shot 2016-11-15 at 7.46.33 pm.png|thumbnail! Blue ocean seems a bit wrong by design here - the artifacts tab can't be a flat listing. It should be loaded on demand, and probably never load the whole set of artifacts at once. |
| Link |
New:
This issue relates to |
| Epic Link | New: JENKINS-35759 [ 171771 ] |
| Description |
Original:
When people archive large numbers of files, blue ocean shows this up as lots of items of json test, vs a heirachy that can be browsed. For example, https://ci.jenkins.io/blue/rest/organizations/jenkins/pipelines/Plugins/pipeline-model-definition-plugin/activities/?start=0&limit=1 Will return 23000 lines of json, much of which is in the artifacts field (nearly 2000 objects, each with 5 or so items). This is as a directory structure has been archived with lots of small files, but blue ocean shows this up as a flat listing. This can not scale and seems to be wrong by design. The classic jenkins provides a browsable view of artifacts: eg !Screen Shot 2016-11-15 at 7.44.41 pm.png|thumbnail! This doesn't fetch ANY artifact data until it is asked for (vs blue ocean that gives you a flat list). The raw json: https://gist.github.com/michaelneale/988cfc3ce44458ec9b094ef5dd97c9b3 The job: https://ci.jenkins.io/job/Plugins/job/pipeline-model-definition-plugin/ The pretty printed json: !Screen Shot 2016-11-15 at 7.46.33 pm.png|thumbnail! Blue ocean seems a bit wrong by design here - the artifacts tab can't be a flat listing. It should be loaded on demand, and probably never load the whole set of artifacts at once. |
New:
When people archive large numbers of files, blue ocean shows this up as lots of json objects, flattened out, vs a heirachy that can be browsed. For example, https://ci.jenkins.io/blue/rest/organizations/jenkins/pipelines/Plugins/pipeline-model-definition-plugin/activities/?start=0&limit=1 Will return 23000 lines of json, much of which is in the artifacts field (nearly 2000 objects, each with 5 or so items). This is as a directory structure has been archived with lots of small files, but blue ocean shows this up as a flat listing. This can not scale and seems to be wrong by design. The classic jenkins provides a browsable view of artifacts: eg !Screen Shot 2016-11-15 at 7.44.41 pm.png|thumbnail! This doesn't fetch ANY artifact data until it is asked for (vs blue ocean that gives you a flat list). The raw json: https://gist.github.com/michaelneale/988cfc3ce44458ec9b094ef5dd97c9b3 The job: https://ci.jenkins.io/job/Plugins/job/pipeline-model-definition-plugin/ The pretty printed json: !Screen Shot 2016-11-15 at 7.46.33 pm.png|thumbnail! Blue ocean seems a bit wrong by design here - the artifacts tab can't be a flat listing. It should be loaded on demand, and probably never load the whole set of artifacts at once. |
| Description |
Original:
When people archive large numbers of files, blue ocean shows this up as lots of json objects, flattened out, vs a heirachy that can be browsed. For example, https://ci.jenkins.io/blue/rest/organizations/jenkins/pipelines/Plugins/pipeline-model-definition-plugin/activities/?start=0&limit=1 Will return 23000 lines of json, much of which is in the artifacts field (nearly 2000 objects, each with 5 or so items). This is as a directory structure has been archived with lots of small files, but blue ocean shows this up as a flat listing. This can not scale and seems to be wrong by design. The classic jenkins provides a browsable view of artifacts: eg !Screen Shot 2016-11-15 at 7.44.41 pm.png|thumbnail! This doesn't fetch ANY artifact data until it is asked for (vs blue ocean that gives you a flat list). The raw json: https://gist.github.com/michaelneale/988cfc3ce44458ec9b094ef5dd97c9b3 The job: https://ci.jenkins.io/job/Plugins/job/pipeline-model-definition-plugin/ The pretty printed json: !Screen Shot 2016-11-15 at 7.46.33 pm.png|thumbnail! Blue ocean seems a bit wrong by design here - the artifacts tab can't be a flat listing. It should be loaded on demand, and probably never load the whole set of artifacts at once. |
New:
When people archive large numbers of files, blue ocean shows this up as lots of json objects, flattened out, vs a heirachy that can be browsed. For example, https://ci.jenkins.io/blue/rest/organizations/jenkins/pipelines/Plugins/pipeline-model-definition-plugin/activities/?start=0&limit=1 Will return 23000 lines of json, much of which is in the artifacts field (nearly 2000 objects, each with 5 or so items). This is as a directory structure has been archived with lots of small files, but blue ocean shows this up as a flat listing. This can not scale and seems to be wrong by design. The classic jenkins provides a browsable view of artifacts: eg !Screen Shot 2016-11-15 at 7.44.41 pm.png|thumbnail! This doesn't fetch ANY artifact data until it is asked for (vs blue ocean that gives you a flat list). The raw json: https://gist.github.com/michaelneale/988cfc3ce44458ec9b094ef5dd97c9b3 The job: https://ci.jenkins.io/job/Plugins/job/pipeline-model-definition-plugin/ The pretty printed json: !Screen Shot 2016-11-15 at 7.46.33 pm.png|thumbnail! Blue ocean seems a bit wrong by design here - the artifacts tab can't be a flat listing. It should be loaded on demand, and probably never load the whole set of artifacts at once. NOTE: it seems that with classic it will list the files but only when it is a small amount. |
Seems like the stash/archive step in pipeline should be creating a zip when it is asked to archive a directory Vs archiving each individual file in the directory, which is what it seems to be doing. If it did that then I think this problem would just go away for BO.
| Link |
New:
This issue is blocking |
tfennelly yeah it would (it may be doing that behind the scenes) but I think we have to support existing core behavior for now - which allows you to browse a tree on the web vs download or an explosion of files 
jamesdumay unless I am mistaken this is a critical design fault. I had never looked at how artifacts had worked before but the way it seems to work is not feasible. This needs to be looked at ASAP, perhaps even removed for now until it can be solved.