Hi, we are using the Splunk_TechnologyAdd-on(TA)_Linux(NIX) for process Monitoring.
Recently I noticed that not all processes are listed. For example when I'm running the ps.sh script from the splunk_ta_nix and count the lines ps.sh |wc -l I've got more than 500 Processes. But in Splunk, i'm only able to find exact 150 Processes. It's the same on every host.
So there are a lot of processes missing, I already opened a case about 3 weeks ago without any progress so far.
Does anyone have the same issue? Or has anyone an idea how to whitelist the only needed process information?
yes, you have to install the TA also on the Indexer, we missed that. Afterwards everthing works fine.
Just as hint, we found it very annoying that you can't filter which processes are really needed and you always have to index every single process. So we created another app which is filtering the really needed processes on the indexer and throw away any unneeded process information to keep down the license usage.
REGEX = .
DEST_KEY = queue
FORMAT = nullQueue
and you have to add the following props within the Splunk_TA_nix
SHOULD_LINEMERGE = false
LINE_BREAKER = ([\r\n]+)
DATETIME_CONFIG = CURRENT
KV_MODE = none
PREAMBLE_REGEX = USER PID PSR pctCPU CPUTIME pctMEM RSZ_KB VSZ_KB TTY S ELAPSED COMMAND ARGS
EXTRACT-fields = (?\w+)\s+(?\d+)\s+(?[\w\?]+)\s+(?[\d.]+)\s+(?[\d.:-]+)\s+(?[\d.]+)\s+(?\d+)\s+(?\d+)\s+(?[\d\w\?\/]+)\s+(?\w+)\s+(?[\d:-]+)\s+(?[\w-.\d\/()]+)\s+(?.*)
Depens on which role your instance is acting. We sepperate the workload, some Server are Searchheads and some are responsible only for Indexing the Data. You can have a Search- and Indexinginstance on one Hardware, but it's not recommanded.