Hello All,
I have a GAE (Java, EU) project, running for several years. Billing enabled and all.
During the last weeks, I am experiencing very strange instance lifecycle behavior.
Instances seem to shutdown completely and then restart. This leads to high latencies and high costs as well - because when they stop and restart, 2-3 instances appear (instead of one) probably due to the high latency.
Check the attached graph. This fluctuation is not normal. And has a direct impact on cost. Sometimes I hit 3x the cost of the good, old times when my workload was processed by one, always on instance.
Are you facing capacity issues in the EU datacenter or something?