fivetran / dbt_jira

Data models for Fivetran's Jira connector built using dbt.
https://fivetran.github.io/dbt_jira/
Apache License 2.0
8 stars 15 forks source link

[Feature] Lookback window for Incremental Models #117

Closed fivetran-joemarkiewicz closed 5 months ago

fivetran-joemarkiewicz commented 8 months ago

Is there an existing feature request for this?

Describe the Feature

Currently the jira__issue_field_history model (and it's upstream models) have significant incremental logic in order to handle the high volume of data when it comes to issue field history. However, the logic is built to assume all issue field updates are synced in order and that has been seen to not be the case. There are instances where issue field history records are synced "late" and may not come in for a day or so. This then results in an incorrect end model as the incremental strategy assumes all records are synced in order.

Therefore, we should apply a lookback window within the incremental logic to ensure records synced late are captured in subsequent incremental runs. This will result in an impact to the performance of the models. However, if we do this right, it will be insignificant and more accurate which is the route we would prefer.

Describe alternatives you've considered

Running a full refresh periodically to capture these late arrival records.

Are you interested in contributing this feature?

Anything else?

No response

fivetran-reneeli commented 8 months ago

For reference

https://docs.getdbt.com/best-practices/materializations/4-incremental-models#late-arriving-facts