Two fixes for slow transition issue:
1. Use deep copy instead of a shallow copy in getting available workers for a particular user. (1571)
2. Use disk_used when checking for disk quota left for each user. (1589)
Other changes include:
1. Check for bundle existence before deduction resources (1588)
2. Add a mechanism to configure worker-manager specific values (1569)
3. Install AWS Elastic Container Registry Login helper on Worker image (1567)
4. Test restarting worker in Travis tests (1544)
5. Support unicode in current frontend (1568)
6. Respect minimum number of workers on worker manager (1557)
7. Add database migration for the shared filesystem workers (1556)
8. Enable shared filesystem workers again (1534)