🚨 🚨 ✨ Source Jira: Migrate IssueComments and IssueWorklogs streams to low code

[tolik0]

What

Migrate IssueComments and IssueWorklogs streams to low code based on the CDK update to add an incremental dependency to the parent stream. Corresponding CDK PR: #38211

How

The IssueComments and IssueWorklogs streams have been migrated to low code. Their parent stream, Issues, will now be read incrementally using the state from the child streams. The parent state will be saved to the child state.

User Impact

This change is not reversible.
This PR introduces a breaking change due to the state format modification. Migration to the new version will be done automatically, but the new state format is incompatible with the previous version. The first sync after migration may be slow as the connector needs to reread all parent records.

Old state format:

"stream_state": {
  "updated": "2222-05-08T03:04:45.056-0700"
}

New state format:

"stream_state": {
  "state": [
    {
        "updated": "2222-04-15T11:40:03.372-0700"
    }
  ],
  "parent_state": {
    "issues": {
      "states": [
        {
          "partition": {
            "parent_slice": {},
            "project_id": "10000"
          },
          "cursor": {
            "updated": "2222-10-12T13:43:50.735-0700"
          }
        },
        {
          "partition": {
            "parent_slice": {},
            "project_id": "10016"
          },
          "cursor": {
            "updated": "2222-07-05T12:57:51.258-0700"
          }
        }
      ]
    }
  }
}

Can this PR be safely reverted and rolled back?

• YES 💚
• NO ❌

[vercel]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

Name	Status	Preview	Comments	Updated (UTC)
airbyte-docs	✅ Ready (Inspect)	Visit Preview	💬 Add feedback	Oct 29, 2024 4:16pm

[girarda]

let's be super clear that this is not revertible

[tolik0]

Updated the message

[tolik0]

Regression tests - https://github.com/airbytehq/airbyte/actions/runs/11554415703/job/32157573020
The value of the cursor changed due to new format:
Previous - 2021-04-15T11:39:46.574000-07:00, New - 2021-04-15T11:39:46.574-07:00

[maxi297]

LGTM! I'm just wondering why we emit so many more state messages in the new version based on the regression tests here

[tolik0]

The difference in state count is due to the Python implementation previously emitting state messages at intervals, whereas now we emit them after every slice.