(litmus-portal): Backend Issues

[Jonsy13]

This issue to created to track backend issues -

• Not getting labels in downloaded manifest, it is getting overridden in backend with workflow_id and cluster_id values.
• With multiple-argo servers, the workflows are getting schedules multiple times.
• SA and Cluster roles creation having an issue. Random behavior.
• Namespaced Scope is having issues. (RBAC issues in event-tracker, Service Accounts etc)
• Workflow is still in running state (scheduled 7 days ago), the workflow pods are not present now. Maybe we can remove these entries from the database after a fixed duration
• Resource limits in Namespaced mode are making the servers slower to respond.
• On first Instantiation, the backend server is getting crashed.

time="2021-08-24T12:03:36Z" level=info msg="Go Version: go1.16.7"
time="2021-08-24T12:03:36Z" level=info msg="Go OS/Arch: linux/amd64"
time="2021-08-24T12:03:36Z" level=info msg="Connected To MongoDB"
panic: runtime error: invalid memory address or nil pointer dereference
[signal SIGSEGV: segmentation violation code=0x1 addr=0x10 pc=0x198d2ec]

goroutine 1 [running]:
main.validateVersion(0x2bf6660, 0x2bf6660)
	/gql-server/server.go:54 +0x12c
main.main()
	/gql-server/server.go:68 +0xfa

• (Edge Case): While re-running a workflow with revert-chaos disabled, we get the chaos-engine already present error. Need to handle this.
• YAML Errors need to handle (probes/labels and chaosEngines) (Discussion: Should be handled in backend or frontend)
  API request can be made to validate the YAML by unmarshaling in backend after saving changes in the editor.
• cleanup script requires the PVC to be deleted or else old data will persist.
• The workflow got completed but the data we were getting from the backend still show the experiments in running phase
• Agents are added in Table, Even when registration failed.