SnapshotShardsService should use TransportMasterNodeAction for shard status updates

[imotov]

While reviewing stuck a stuck snapshot issue related to #25863 (comment), @ywelsch pointed out that we can improve reliability of shard updates by switching the mechanism to TransportMasterNodeAction (we are currently using a plain TransportRequestHandler that doesn't have a robust retry mechanism of TransportMasterNodeAction). This is a simple change that can significantly improve reliability of snapshot on flakey networks.

[jyoti0208]

I would like to work on this.

[ywelsch]

@jyoti0208 This is already assigned and there is already an open PR for this - no need to work on this.

[jyoti0208]

Is there any other task which I can take as the first-time contributor for this project.

[ywelsch]

Sure, any issue marked as adoptme is up for grabs. Issues that are additionally marked as low hanging fruit should be simpler tasks.

Please also have a look at our contributing guide which will help you get started.