Open emilburzo opened 5 years ago
Could it be that 3 of the previous instances were not longer in the running state and maybe already terminating due to a scale in event?
That's very unlikely, traffic was in a (slowly) increasing trend
Take 2: for some reason both the 2355 and 2345 ASGs still existed and our code took the capacity from the 2345 instead of the 2355 for the new 2359 ASG
Very interesting, is it possible to work-around such situations?
This occurred again today, any updates on this?
Yeah, I guess there was a failed earlier deploy. Any chance to get that fixed? By e.g. taking the currently registered ASG (assuming that there is only one per ELB)? Alternatively if both ASGs where still registered: The sum of all instances which would be safer than just the instances of one ASG
Hello,
We just had a situation where there were 5 old running instances, but the deploy only started 2 new instances (which wasn't enough to handle the current traffic)
What could have caused this?