Smart dynamic restore for Kubernetes based applications
One example method includes discovering an application topology, discovering relationships between microservices that are elements of the application topology, ranking the microservices, identifying one or more of the microservices as a persistency microservice, selecting one or more persistency microservices for backup, and defining a backup policy based on the microservice relationships, the microservice rankings, and the selected persistent microservices. A backup operation is then performed that includes backing up persistent data and/or persistent metadata generated and/or modified by one or more of the persistency microservices.
1. A non-transitory storage medium having stored therein instructions which are executable by one or more hardware processors to perform operations comprising:
discovering a topology of a microservice;
discovering a topology of an application, and discovering the application topology comprises identifying a group of microservices, including the microservice, that are components of the application;
discovering relationships between the microservices in the group of microservices that were identified during the discovering of the application topology, and discovering relationships between the microservices comprises discovering whether a microservice in the group of microservices is dependent upon, or independent of, another microservice in the group of microservices;
identifying a sub-group of persistency microservices in the group of microservices;
ranking the persistency microservices of the sub-group of persistency microservices;
selecting one or more of the ranked persistency microservices for backup; and
defining a backup policy based on the microservice relationships, the persistency microservice rankings, and the selected persistency microservices, and the backup policy, when executed, causes backup of the one or more persistency microservices.
2. The non-transitory storage medium as recited in claim 1 , wherein the persistency microservices are ranked according to their importance relative to each other.
3. The non-transitory storage medium as recited in claim 1 , wherein, in operation, one of the persistency microservices generates new and/or modified data and/or metadata that is persistently stored.
4. The non-transitory storage medium as recited in claim 1 , wherein discovery of the application topology is performed with a platform application program interface (API).
5. The non-transitory storage medium as recited in claim 1 , wherein the platform API is a Kubernetes API.
6. The non-transitory storage medium as recited in claim 1 , wherein discovery of relationships between microservices is performed with a service mesh.
7. The non-transitory storage medium as recited in claim 1 , wherein the application topology is a topology of a cloud native application.
8. The non-transitory storage medium as recited in claim 1 , wherein the operations further comprise performing a backup operation based on the backup policy.
9. The non-transitory storage medium as recited in claim 8 , wherein performing the backup operation comprises backing up persistent data and/or persistent metadata generated and/or modified by one or more of the persistency microservices.
10. The non-transitory storage medium as recited in claim 1 , wherein the operations further comprise:
detecting respective rankings of the persistency microservices; and
defining a restore policy concerning the microservices, wherein defining the restore policy comprises:
specifying a restore order for the microservices, and a restore order for the persistency microservices is specified based on the respective rankings of the persistency microservices;
identifying microservices that are candidates for restoration; and
specifying timing of a restore for each of the candidates.
11. The non-transitory storage medium as recited in claim 10 , wherein the persistency microservices are ranked according to their importance relative to each other, and the operations further comprise restoring the candidates in order of their rank.
12. The non-transitory storage medium as recited in claim 10 , wherein each of the restore candidates is a respective one of the persistency microservices, and each of the persistency microservices is associated with respective persistent data and/or persistent metadata.
13. The non-transitory storage medium as recited in claim 10 , wherein the operations further comprise restoring the restore candidates according to the restore policy.
14. The non-transitory storage medium as recited in claim 13 , wherein restoring the restore candidates comprises restoring a failed persistency microservice before the failed persistency microservice impairs operation of an application of which that persistency microservice is an element.
15. The non-transitory storage medium as recited in claim 13 , wherein restoring the restore candidates comprises restoring multiple microservices together.
16. The non-transitory storage medium as recited in claim 13 , wherein restoring the restore candidates comprises restoring one of the restore candidates directly from secondary storage.
17. The non-transitory storage medium as recited in claim 13 , wherein as between first and second restore candidates, the first restore candidate is more likely to be used by an application before the second restore candidate is likely to be used by the application, and restoring the restore candidates comprises restoring the first restore candidate before restoring the second restore candidate.
18. The non-transitory storage medium as recited in claim 13 , wherein restoring the restore candidates comprises restoring less than all of the microservices.
19. The non-transitory storage medium as recited in claim 13 , wherein one of the restore candidates is run from protection storage prior to being restored to primary storage.