Pengfei Ni

@JoelSpeed Good catch and thanks for reporting the issue. This is indeed a bug we should fix.

@MartinForReal has merged #5652 as first step. For next step, we should add timeouts in context. @MartinForReal could you work a fix on this?

this is still required as multiple Azure resource (NSG, LB, public IP, VMSS VM) may be updated for a service, and each of would have ETAG to guard the consistency....

/milestone v1.27

Thanks for the discussion and conclusion! /close

/assign @MartinForReal

This would be done together with Azure track2 SDK migration.

> I think using specific timeouts throughout the codebase would be appropriate as it would allow you to asses what timeout you need for each specific use case, rather than...

> Without this, ConcurrentNodeSyncs is always 0, as a result of which the node controller does not work. @vitaliy-sn could you share more details of the issue? would you like...

Could you bump the versions to latest stable version v0.29.0?