Turns Out AI Is Not Good at Database Transaction Scheduling (adrs-ucb.notion.site)

We modelled using Redshift scheduling automation, that suggested a welcome 20% improvement.

We applied it and it run 20% slower.

As it runs it is supposed to learn how to optimise scheduling. Not sure how it gets over the burst at month, quarter, and year end.

turns out DB transactions are a hard problem...

I wonder what the optimal for these workloads are?

For the offline problem, AI found a new solution that was 34% faster. Interesting results!

What about other use cases

Seems like AI hype

Transactions seem like a hard problem for scheduling, wonder how the authors got this to work in the original paper?