In the existing PR Review model the feature flags settings are reused from the existing Staging environment. This can cause issues when new feature flags need to be tested on PR Review branches as they cannot be turned on/off independently of the Staging Environment.
Existing solution (best viewed as GitHub issue):
flowchart TD
A[Staging Image Startup] -->|Check Flags| B{Flags Exist}
F[PR Review App] --> |read / write| E
G[Staging App] --> |read / write| E
B --> |yes| C[Update new / removed Flags]
B --> |no| D[Create Flags\n with default values ]
C --> E[Redis 'flags' contained in set]
D --> E
A solution needs to be developed to allow PR Review environments to run in isolation of the feature flag settings.
There is an existing Environment variable in PR Review containers that identify them as such. The REVIEW_ENV variable is only set in the PR Review image.
Acceptance Criteria:
[ ] The solution must not rely on code changes that can be accidentally merged into develop and overwrite staging / production settings
[ ] The solution must clean up after itself if feature flag artifacts are created during it's lifetime
[ ] The solution should be transparent to the developer working on new feature flags. No new process other then modifying the existing flag-initialization json should be implemented.
In the existing PR Review model the feature flags settings are reused from the existing Staging environment. This can cause issues when new feature flags need to be tested on PR Review branches as they cannot be turned on/off independently of the Staging Environment.
Existing solution (best viewed as GitHub issue):
A solution needs to be developed to allow PR Review environments to run in isolation of the feature flag settings. There is an existing Environment variable in PR Review containers that identify them as such. The
REVIEW_ENV
variable is only set in the PR Review image.Acceptance Criteria:
develop
and overwrite staging / production settings