We are planning to do some linux patching on some of our Prod servers and were thinking of setting some splunk alert which detects and reports any kind of anomalies in our logs after the patching is done. However, I have had varied results with anomalousvalues and anomalydetection. Currently I am using a combination of both. My concern however is, suppose we have 4 hosts under one sourcetype, and all the hosts restart after patching and this is logged in my logs, would splunk treat it as anomaly of each host or club them together and not treat it as an anomaly. I guess, the gist of my question is, can I have anomalydetection for each host without setting an alert for each individual host?
sourcetype="prod_gw" source="/opt/mqm/logs/err01.log" | anomalousvalue |anomalydetection is my current search for alert where sourcetype is a group of 4 hosts.