Question
We are attempting to implement DORA metrics via Apache DevLake for a large-scale architecture consisting of roughly 1,000 individual deployable units.
Currently, our Jira setup makes it incredibly difficult to isolate data per deployable unit using the standard DevLake UI workflow.
Our Architecture & Problem:
Jira Project Topology: We have a single monolithic Jira project used to cover a product feature such as installations that tracks all of these deployable units. so one project can track 50+ deployable units. In devlake each project is a deployable unit.
Board Chaos: Within this project, there are numerous boards. These boards are dynamic, unstandardized, and multi-purpose (some look at Epics, some look at Bugs, some use Kanban, some use Scrum). There are no strict rules.
The Data Pollution Issue: Multiple deployable units live on the same boards. If we map a shared Jira board to a specific DevLake project (e.g., a DevLake project dedicated to abc.ui.host), DevLake pulls all issues on that board. This heavily pollutes our DORA metrics (Change Failure Rate, Lead Time for Changes) with data from unrelated services like abc.lambda.
Scale Bottleneck: The recommended workaround of creating dedicated Jira boards filtered by JQL (e.g., project = ABC AND component = abc.ui.host) does not scale for 1,000 deployable units, especially when historical data lacks perfect component tagging.
Questions for the Maintainers:
Is there an unexposed way to inject JQL filters per Data Scope? Can we bypass the UI and alter the blueprint JSON directly via the DevLake API to append a custom JQL query (e.g., filtering by a dynamic field, custom label, or component) during the collection phase?
How does DevLake recommend handling multi-tenant/multi-service Jira boards? If 50 different microservices are being tracked on a single team Kanban board, how do we isolate DORA metrics per microservice without creating 50 distinct boards in Jira?
Future Roadmap: If this is not currently supported, is there an active design or plan to introduce an inline JQL text filter directly inside the Data Scope or Scope Config UI to allow granular issue filtering?
Question
We are attempting to implement DORA metrics via Apache DevLake for a large-scale architecture consisting of roughly 1,000 individual deployable units.
Currently, our Jira setup makes it incredibly difficult to isolate data per deployable unit using the standard DevLake UI workflow.
Our Architecture & Problem:
Jira Project Topology: We have a single monolithic Jira project used to cover a product feature such as
installationsthat tracks all of these deployable units. so one project can track 50+ deployable units. In devlake each project is a deployable unit.Board Chaos: Within this project, there are numerous boards. These boards are dynamic, unstandardized, and multi-purpose (some look at Epics, some look at Bugs, some use Kanban, some use Scrum). There are no strict rules.
The Data Pollution Issue: Multiple deployable units live on the same boards. If we map a shared Jira board to a specific DevLake project (e.g., a DevLake project dedicated to abc.ui.host), DevLake pulls all issues on that board. This heavily pollutes our DORA metrics (Change Failure Rate, Lead Time for Changes) with data from unrelated services like abc.lambda.
Scale Bottleneck: The recommended workaround of creating dedicated Jira boards filtered by JQL (e.g., project = ABC AND component = abc.ui.host) does not scale for 1,000 deployable units, especially when historical data lacks perfect component tagging.
Questions for the Maintainers:
Is there an unexposed way to inject JQL filters per Data Scope? Can we bypass the UI and alter the blueprint JSON directly via the DevLake API to append a custom JQL query (e.g., filtering by a dynamic field, custom label, or component) during the collection phase?
How does DevLake recommend handling multi-tenant/multi-service Jira boards? If 50 different microservices are being tracked on a single team Kanban board, how do we isolate DORA metrics per microservice without creating 50 distinct boards in Jira?
Future Roadmap: If this is not currently supported, is there an active design or plan to introduce an inline JQL text filter directly inside the Data Scope or Scope Config UI to allow granular issue filtering?