Spotinst is not waiting for in-flight requests to finish if drainingTimeout is not set, which can cause spikes of ELB errors when Elastigroups scale down. Suggestion is to make senza use a sane default (like 120s) to avoid issues during down-scaling of stacks from Spotinst.
otrosien
commented
3 years ago
Spotinst is not waiting for in-flight requests to finish if drainingTimeout is not set, which can cause spikes of ELB errors when Elastigroups scale down. Suggestion is to make senza use a sane default (like 120s) to avoid issues during down-scaling of stacks from Spotinst.