We currently use in house load gens that are VMs with the following spec:
Intel(R) Xeon(R) CPU E5-2620 v3 @ 2.40GHz, 4 cores
Virtual memory 9.75 GB total
Real memory 7.45 GB total
We have recently seen issues when running load where a load gen will drain from the running composition, due to breaching the Monitor.LoadAvg.XMinuteThreshold(s).
The default settings for LoadAvg (per core) is:
- 1 Min - 50
- 2 Min - 25
- 5 Min – 15
- 15 Min - 5
Unless I missunderstand the figures (4 cores = 400, so 100 per core), they do seem a bit on the low side. Would we face issues where maestros are overloaded if we were to increase these figures? How much leeway do we have before results are affected?