feels like moving problem from operations to developer. with staging: keeping almost identical environment with production, syncing data if necessary to test l

feels like moving problem from operations to developer. with staging: keeping almost identical environment with production, syncing data if necessary to test load. can be considered as mostly Ops related work. without staging: of course feature flags should be used when necessary, but if there is no staging environment we should create feature flag for almost every pull request, which in turn create a lot of legacy which should be deleted later. add nested feature flags into this and now you have problem of seeing which conditions are executed when multiple teams are working in parallel