We need to double check this with SD team cc @riuvshin but we currently set 2.4GB as mem quota limit even if there are 3GB available in the Che namespace (that's because before we run the wsmaster too in the che namespace).
The fix shold include both Che tenant namespace definition yaml and rh-che OpenShift template too.
@l0rd can you please test this on stg? I mean if 3g workspaces are working properly because from what I know it can reach some RAM/CPU limits.
Also please make sure SD team is aware if you are going to change limits because currently we are assuming that workspaces takes 2.4g for capacity clusters capacity calculations.
@riuvshin this is not in current sprint backlog. Hence it's unlikely that we can test it before the end of current sprint.
@ibuziuk @l0rd here is the issue https://gitlab.cee.redhat.com/dtsd/housekeeping/issues/794 where we had issues with increasing ws ram limit to 2.5g
@riuvshin PR has been sent - https://github.com/fabric8-services/fabric8-tenant-che/pull/104
@riuvshin could you please take care of https://gitlab.cee.redhat.com/dtsd/housekeeping/issues/2142 ? We need just to update cm for now - no need to rollout latest deployment
PR for updating quotas on prod - https://github.com/openshiftio/saas-openshiftio/pull/982
@riuvshin once this is merged we would need to proceed with massive the massive tenant updates on prod & prod-preview. In general I would propose the following sequence of actions:
rhche cm with new values for workspace ram https://gitlab.cee.redhat.com/dtsd/housekeeping/issues/2142WDYT ?
BTW, I have tested new quotas on dev cluster against prod-preview and was able to create workspace successfully with 3G allocated to the dev machine:

limits updated on preview and prod
New limits are now available on prod / prod-preview. Closing