[kirmerzlikin]

The approaches described in the article are clever and the results are impressive, so kudos to the team!

However, while reading, it felt weird that they treated Skipper as a third-party dependency, even though it is developed and operated by their colleagues. I mean, they could've tried contributing some updates and improvements to Skipper, maybe reusing Skipper's discovery mechanism, instead of rolling out their own, completely independent LB module.

And yes, it's true that in big companies it can be hard to organise cross-team collaboration, but IMO it's still worth a try.

[43920]

I had a similar thought - it seems like most of the concerns could be addressed by adding better request tracing and improving the resiliency of the existing system. I wonder if this is just the result of an organizational tension where the load balancing team isn’t motivated to address this particular use case (which seems likely if the organization is large enough).