Hi Guys,
I have a very large dynamic inventory with ~ 6000 hosts.
Does anyone have any experience with playbook runs across this many hosts? I’m running into issues (frequent and long CPU spikes, significant memory usage, oom kills, etc) when running very simple playbooks across all of my hosts.
If I split my inventory into 500 node batches, I can initiate multiple runs and progress through all hosts cleanly in about 15 minutes.
Can obviously provide more details but just curious as to whether anyone has any experience or suggestions when running across very large numbers of systems?
Thanks - Dan.