Hi Jordan,
Please see inline for my responses:
"I see in the doc that the file should rarely exceed 2 GB yet almost all of ours are in the 4.5 GB range."
<Kulbir>It's not atypical to have this huge size baselines.db but since it's a file based database very often it could get corrupted easily(EM crash\improper shutdown) and that could also cause size to grow pretty big.Typically corruption would be accompanied by error\warning messages in EM logs in which case it's recommended to delete the original baselines.db and restart EM.
Baselines would subsequently be recalculated by EM as application data comes in over a period of next few days.
My questions: Can this file size impact performance?
<Kulbir.> Yes a large baselines.db file can impact EM performance especially if it's already corrupted.
"Is there a way to clean it up without loosing all of our baseline data? "
<Kulbir>No you have to delete it if it's corrupted, but unlike Smartstor this doesn't store any real metric performance data, only derived metrics and will be rebuilt over period of next few days.
See https://docops.ca.com/display/APMDEVOPS97/The+Metric+Browser+Tab#TheMetricBrowserTab-HeuristicsandMetricBaselines
Does the baseline.db file have any age limit settings?
<No>
Please note that baselines.db has been replaced by variance.db in APM 10 and later releases ,which is part of Differential Analysis feature released in APM 10.
It does far more than simple baselining and you can read more details here:
https://docops.ca.com/ca-apm/10/en/ca-apm-release-notes#CAAPMReleaseNotes-DifferentialAnalysis
Hope this helps.
Regards,
Kulbir.