When you have a container with both feed and query, you risk that during warmup the throughput optimized

feed pipeline will overwhelm the search pipeline during warmup.
Here we add an option to enforce a soft start of the feed pipeline, by gradually increasing max number of feed threads.
It might not be the perfect solution as mesages will pile up in the dynamic throttle policy, but that should have a builtin
softstart already. So I think this might a good fit.

0 files changed, 0 insertions, 0 deletions