You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Description
As a move towards improving our current deployment processes, it would be good to investigate the feasibility of a basic Continous Delivery (CD) system for the automatic deployment of a designated work-in-progress branch as PRs are merged in. At this stage the aim would be to have a working copy of the latest commit from primary active branch for repos available for testing. This could be for use by developers, users, or integration-suite testing.
Acceptance Criteria
The following outputs from the research are produced:
Feedback from the team on what would be a comfortable and appropriate system
A sketched-out process on how the automatic deployments might be implemented. This could later be turned into stories for implementation
Possibly a proof-of-concept for demonstration purposes
Who the primary contacts are for this work @StephenHulme, PSD team (most likely principles in practise)
Knowledge or Stake holders
PSD Team
Stakeholders who would interact with the deployed applications.
Additional context or information
Could be linked with #450
The text was updated successfully, but these errors were encountered:
I think an automated deployment will make collaboration for the use of UAT environment very difficult. Identifying WIP branches is also particularly difficult (I am taking WIP branch to mean feature flagged branch). Handling EPIC stories, branches, story, PR and multiple branch dependencies and relationships are very difficult too.
I agree. My thinking here is more of a roadmap to an achievable system setup, rather than simply naively automating a subset of out current process.
I don't want to prescribe a particular method or outcome in the story, but one thought I had was to investigate what it might look like if we split UAT into two identical 'physical' setups, but one with an automated deployment for testing int-suite against, and the other similar to now where it is generally aimed towards end-user testing.
A workflow where one could merge a change into develop and gets notified of int-suite passing/failing within an hour could be a powerfull tool for developers, particularly for dependency updates. If it fails, you get a notification and can revert the change soon afterwards - no harm done and much information gained for little effort.
I have effectively (and somewhat unintentionally) created a ticket outlining the SIT (System Integration Test) envionment outlined in the PSD Strategy 2024 Q3 Update, Reduce UAT Contention section.
Description
As a move towards improving our current deployment processes, it would be good to investigate the feasibility of a basic Continous Delivery (CD) system for the automatic deployment of a designated work-in-progress branch as PRs are merged in. At this stage the aim would be to have a working copy of the latest commit from primary active branch for repos available for testing. This could be for use by developers, users, or integration-suite testing.
Acceptance Criteria
The following outputs from the research are produced:
Who the primary contacts are for this work
@StephenHulme, PSD team (most likely principles in practise)
Knowledge or Stake holders
Additional context or information
Could be linked with #450
The text was updated successfully, but these errors were encountered: