Currently we have a weekly deployment job that discovers, builds and deploys stress tests across all repositories. However we do not have any CI in the language+service specific pipelines, which can cause breakages in the stress tests to be discovered too late (e.g. shifting dependencies causing failing docker image builds). We should add logic to the language CI pipelines to run our stress build scripts.
This will also require updating our stress deploy script to have a build only flag and/or WhatIf support (which arguably we have needed for a while).
benbp
commented
2 years ago
Currently we have a weekly deployment job that discovers, builds and deploys stress tests across all repositories. However we do not have any CI in the language+service specific pipelines, which can cause breakages in the stress tests to be discovered too late (e.g. shifting dependencies causing failing docker image builds). We should add logic to the language CI pipelines to run our stress build scripts.
This will also require updating our stress deploy script to have a build only flag and/or WhatIf support (which arguably we have needed for a while).