Thank you for considering Dgraph. While it is technically possible to do so, it is generally not recommended as it can lead to scalability and maintainability issues in the long run. It is better to have separate pods for each application, so they can be managed independently and scaled as needed. I hope this helps clarify the best practice for deploying applications on Kubernetes.
You could use the dgraph/standalone image which contains a Zero and an Alpha instance running in the same container. However, clustering is not possible with this approach. Perhaps, if you create a standalone script that somehow manages the entry and exit of Dgraph nodes as you scale your cluster, it could work. But, it seems like a complex solution to maintain. I would not recommend it.