My log files log a bunch of messages in the same instance, so simply search for a message id followed by a count will not work (I will only count 1 per event when I want to count as many as 50 per event). I want to first narrow down my search to the events which show messages being sent ("enqueued"), and then count all instances of the string "mid".
Any ideas? I am very bad with splunk. How to I get all instances of "mid" to be a countable field?
index=* service=myservice "enqueued" "mid" | stats count mid
Your current search doesn't work because you (probably) don't have a field called 'mid'.
To search for strings within the event you can use rex. Try this.
index=* service=myservice "enqueued" "mid"
| rex max_match=0 "(?<mids>mid)"
| eval midCount=mvcount(mids)
| table midCount
BTW, "index=*" is a bad practice. It forces Splunk to search in every index, which really slows things down. After your first search you should know and use the real index name.
Related
I have 5 different searches I am doing in Splunk where I am getting the count of how many results from that search query.
I've had a look at this thread here:
https://answers.splunk.com/answers/757081/pie-chart-with-count-from-different-search-criteri.html
but its not quite working for me, I'm not 100% sure if its what I want.
My search queries all look something like this:
index=A variable="foo" message="Created*" | stats count
index=A variable="foo" message="Deleted*" | stats count
I ideally want to assign each query to a keyword - such as created, deleted, etc, then do a pie chart based on the counts.
The following should be sufficient.
index=A variable="foo" message="Created*" OR message="Deleted*" OR message="<repeat this for any other message types you want>" | stats count by message
If you can provide some more examples of the events you are trying to chart, there may be alternate approaches that can work for you.
This version will extract the key part of the message (Created, Deleted. etc...) into a field called mtype and you can then perform stats on that field.
index=A variable="foo" message="Created*" OR message="Deleted*" OR message="<repeat this for any other message types you want>" | rex field=message "(?<mtype>Created|Deleteted|...)" | stats count by mtype
From Splunk, I am trying to get the user, saved search name and last time a query ran ?
A single Splunk query will be nice.
I am very new to Splunk and I have tried these queries :-
index=_audit action=search info=granted search=*
| search IsNotNull(savedsearch_name) user!="splunk-system-user"
| table user savedserach_name user search _time
The above query , is always empty for savesearch_name.
Splunk's audit log leaves a bit to be desired. For better results, search the internal index.
index=_internal savedsearch_name=* NOT user="splunk-system-user"
| table user savedsearch_name _time
You won't see the search query, however. For that, use REST.
| rest /services/saved/searches | fields title search
Combine them something like this (there may be other ways)
index=_internal savedsearch_name=* NOT user="splunk-system-user"
| fields user savedsearch_name _time
| join savedsearch_name [| rest /services/saved/searches
| fields title search | rename title as savedsearch_name]
| table user savedsearch_name search _time
Note that you have a typo in your query. "savedserach_name" should be "savedsearch_name".
But I also recommend a free app that has a dedicated search tool for this purpose.
https://splunkbase.splunk.com/app/6449/
Specifically the "user activity" view within that app.
Why it's a complex problem - part of the puzzle is in the audit log's info="granted" event, another part is in the audit log's info="completed" event, even more of it is over in the introspection index. You need those three stitched together, and the auditlog is plagued with parsing problems and autokv compounds the problem by extracting all of fields from the SPL itself.
That User Activity view will do all of this for you, sidestep pretty thorny autokv problems in the audit data, and not just give you all of this per search, but also present stats and rollups by user, app, dashboard, even by sourcetypes-that-were-actually-searched
it also has a macro called "calculate pain" that will score a "pain" number for each search, and then sum up all the "pain" in the by-user, by-app, by-sourcetype rollups etc. So that admins can try and pick off the worst offenders first.
it's up on SB here and approved for both Cloud and onprem - https://splunkbase.splunk.com/app/6449/
(and there's a #sideview_ui channel for it in the community slack.)
I want to print the value of a certain field from a set of events that results from running a particular search query. Here's my query:
index=abc "all events that contain this string" sourcetype=prd
Now, this returns certain events that contain a field called traceId. What I want is to extract unique traceIds from the result and print them. Here's the query that I am using currently, but to no avail:
index=abc "all events that contain this string" sourcetype=prd | rex field=_raw "traceId: (?<traceId>.*)"
This query prints all the fields in the event (events are printed as JSON docs.).
Can someone help me with this? I have never worked with Splunk before, so please go easy if the question looks a bit easy.
Thanks!
Answering this without some sample data is almost impossible... still, I think you are getting all data because:
you are not using the fields command to filter your fields of interest. It would go like so: `index=abc "all events that contain this string" sourcetype=prd | rex field=_raw "traceId: (?.*) | fields fiel1, field2, traceId"
your regular expression is greedy, which means traceId field will contain all text from that point to the end of the event. Try to be more specific i.e. \d+ for numeric data or even [^\s]+ for non-blanks.
~HTH
This might be a really simple question, but I haven't been able to find an answer as of yet. I have some raw data from some events that is for example "(duration 5555ms)" and I want to put that in a "| timechart span=1m count by duration" to create a chart that shows when these events took place and their total duration. There is currently no field set up for duration, it is just raw data. How would I get those numbers into my time chart?
You will first need to extract the value for duration into a field. You will most likely use the regex (rex) function for this.
The exact command you need will depend a lot on your data. But for your example "(duration 5555ms)", this should work assuming the value is always in ms.
| rex field=_raw "\(duration (?<duration>\d+)ms.*"
I am looking to see how many servers are reporting into splunk over time. This is a query similar to the one I have tried:
sourcetype=defined | dedup host | timechart count by pop
What is happening is the host gets deduped before the time chart (obviously) so I'm not exactly getting the results I'm looking for.
How can I deduplicate the server list per time slice in the timechart?
Please let me know if further clarification is necessary.
Looks like the option I was looking for was the distinct_count function for charts. This is the final query that is returning results that I am looking for:
sourcetype=defined | timechart dc(host) by pop