I am trying to get a count from transactional information that is retained within raw data in splunk. I have 3-5 transactions that occur.
One has raw data stating: pin match for id 12345678-1234-1234-abcd-12345678abcd or pin mismatched for id etc.
I'm trying to count the number of times the pin match occurs within the transaction time window of 180sec.
I was trying to do something like:
|eval raw=_raw |search index=transa
|eval pinc= if((raw like "%pin match%"),1,0) |stats count(pinc) as Pincount by ID
The issue I'm having is it is counting cumulatively over whatever time I am looking at those transactions. Is there a way to attach it to the ID that is within the message or have it count every one that occurs within that time window?
Thanks!
Presuming the pin status and ID have not been extracted:
index=ndx sourcetype=srctp "pin" "match" OR "mismatched"
| rex field=_raw "pin (?<pin_status>\w+)"
| rex field=_raw "id (?<id>\S+)"
| eval status_time=pin_status+"|"+_time
| stats earliest(status_time) as beginning latest(status_time) as ending by id
| eval beginning=split(beginning,"|"), ending=split(ending,"|")
| eval begining=mvindex(beginning,-1), ending=mvindex(ending,-1)
| table id beginning ending
| sort 0 id
| eval beginning=strftime(beginning,"%c"), ending=strftime(ending,"%c")
After extracting the status ("match" or "mismatched") and the id, append the individual event's _time to the end of the status - we'll pull that value back out after statsing
Using stats, find the earliest and latest status_time entries (fields just created on the previous line) by id, saving them into new fields beginning and ending
Next, split() beginning and ending on the pipe we added to separate the status from the timestamp into a multivalue field
Then assign the last item from the multivalue field (which we know is the timestamp) into itself (because we know that the earliest entry for a status_time should always be "match", and the latest entry for a status_time should always be "mismatched")
Lastly, table the id and time stamps, sort by id, and format the timestamp into something human readable (strftime takes many arguments, %c just happens to be quick)
Related
I have multiple log messages each containing a list of JobIds -
IE -
1. `{"JobIds":["661ce07c-b5f3-4b37-8b4c-a0b76d890039","db7a18ae-ea59-4987-87d5-c80adefa4475"]}`
2. `{"JobIds":["661ce07c-b5f3-4b37-8b4c-a0b76d890040","db7a18ae-ea59-4987-87d5-c80adefa4489"]}`
3. `{"JobIds":["661ce07c-b5f3-4b37-8b4c-a0b76d890070"]}`
I have a rex to get those jobIds. Next I want to count the number of jobIds
My query looks like this -
| rex field=message "\"(?<job_ids>(?:\w+-\w+-\w+-\w+-\w+)+),?\""
| stats count(job_ids)
But this will only give me a count of 3 when I am looking for 5. How can I get a count of all jobIds? I am not sure if this is a splunk limitation or I am missing something in my regex.
Here is my regex - https://regex101.com/r/vqlq5j/1
Also with max-match=0 but with mvcount() instead of mvexpand():
| makeresults count=3 | streamstats count
| eval message=case(count=1, "{\"JobIds\":[\"a1a2a2-b23-b34-d4d4d4\", \"x1a2a2-y23-y34-z4z4z4\"]}", count=2, "{\"JobIds\":[\"a1a9a9-b93-b04-d4d4d4\", \"x1a9a9-y93-y34-z4z4z4\"]}", count=3, "{\"JobIds\":[\"a1a9a9-b93-b04-d14d14d14\"]}")
``` above is test data setup ```
``` below is the actual query ```
| rex field=message max_match=0 "\"(?<id>[\w\d]+\-[\w\d]+\-[\w\d]+\-[\w\d]+\")"
| eval cnt=mvcount(id)
| stats sum(cnt)
In Splunk, to capture multiple matches from a single event, you need to add max_match=0 to your rex, per docs.Splunk
But to get them then separated into a singlevalue field from the [potential] multivalue field job_ids that you made, you need to mvxepand or similar
So this should get you closer:
| rex field=message max_match=0 "\"(?<job_id>(?:\w+-\w+-\w+-\w+-\w+)+),?\""
| mvexpand job_id
| stats dc(job_id)
I also changed from count to dc, as it seems you're looking for a unique count of job IDs, and not just a count of how many in total you've seen
Note: if this is JSON data (and not JSON-inside-JSON) coming into Splunk, and the sourcetype is configured correctly, you shouldn't have to manually extract the multivalue field, as Splunk will do it automatically
Do you have a full set of sample data (a few entire events) you can share?
How to make a query to find the number of occurrences of a string in each event, that is, if a tag occurs more than once in an event, the search should show the number of such tags in each individual event
To find the number of occurrences of a specific string, extract the string, count the number of times it appears in each event, then add those numbers.
index=foo "bar"
| rex max_match=0 "(?<bars>bar")
| eval bar_count=mvcount(bars)
| stats sum(bar_count) as total_bars
My use case is analysing ticket in order to attribute a state regarding all the status of a specific ticket.
Raw data look like this :
Id
Version
Status
Event Time
0001
1
New
2021-01-07T09:14:00Z
0001
1
Completed - Action Performed
2021-01-07T09:38:00Z
Data looks like this after transaction command:
Id
Version
Status
Event Time
state
0001, 0001
1, 1
New, Completed - Action Performed
2021-01-07T09:14:00Z, 2021-01-07T09:38:00Z
Acknowlegdement, Work
I'm using transcation command in order to calculate the duration of acknnowlegdement and resolution of the ticket.
I have predefine rule to choose the correct state. This rules compare the n-1 status (New), and the current status (Completed - Action Performed) to choose the state.
Issue
Each ticket has a different number of status. We can not know in advance the max status number. I can not write a static search comparing each value of the Status field.
Expected Solution
I have a field that inform me the number of index on the status (number of status of a ticket) field.
I want to use a loop (Why not a loop for), to iterate on each index of the field Status and compare the value i-1 and i.
I can not find how to do this. Is this possible ?
Thank you
Update to reflect more details
Here's a method with streamstats that should get you towards an answer:
index=ndx sourcetype=srctp Id=* Version=* Status=* EventTime=* state=*
| eval phash=sha256(Version.Status)
| sort 0 _time
| streamstats current=f last(phash) as chash by Id state
| fillnull value="noprev"
| eval changed=if(chash!=phash OR chash="noprev","true","false")
| search NOT changed="false"
| table *
original answer
Something like the following should work to get the most-recent status:
index=ndx sourcetype=srctp Id=* Version=* Status=* EventTime=* state=*
| stats latest(Status) as Status latest(Version) as Version latest(state) state latest(EventTime) as "Event Time" by Id
edit in light of mentioning g the transaction command
Don't use transaction unless you really really really need to.
99% of the time, stats will accomplish what transaction does faster and more efficiently.
For example:
index=ndx sourcetype=srctp Id=* Version=* Status=* EventTime=* state=*
| stats earliest(Status) as eStatus latest(Status) as lStatus earliest(Version) as eVersion latest(Version) as lVersion earliest(status) as estate latest(state) lstate earliest(EventTime) as Opened latest(EventTime) as MostRecent by Id
Will yield a table you can then manipulate further with eval and such. Eg (presuming the time format is subtractable (ie still in Unix epoch format)):
| eval ticketAge=MostRecent-Opened
| eval Versions=eVersion+" - "+lVersion
| eval Statuses=eStatus+" - "+lStatus
| eval State=estate+", ",lstate
| eval Opened=strftime(Opened,"%c"), MostRecent=strftime(MostRecent,"%c")
| eval D=if(ticketAge>86400,round(ticketAge/86400),0)
| eval ticketAge=if(D>0,round(ticketAge-(D*86400)),ticketAge)
| eval H=if(ticketAge>3600,round(ticketAge/3600),0)
| eval ticketAge=if(H>0,round(ticketAge-(H*3600)),ticketAge)
| eval M=if(ticketAge>60,round(ticketAge/60),0)
| eval ticketAge=if(M>0,round(ticketAge-(M*60)),ticketAge)
| rename ticketAge as S
| eval Age=D+" days "+H+" hours"+M+" minutes"+S+" seconds"
| table Id Versions Statuses Opened MostRecent State Age
| rename MostRecent as "Most Recent"
Note: I may have gotten the conversion from raw seconds into days, hours, minutes, seconds off - but it should be close
As part of my requirements, I have to calculate the duration between two different logs using Splunk query.
For example:
Log 2:
2020-04-22 13:12 ADD request received ID : 123
Log 1 :
2020-04-22 12:12 REMOVE request received ID : 122
The common String between two logs is " request received ID :" and unique strings between two logs are "ADD", "REMOVE". And the expected output duration is 1 hour.
Any help would be appreciated. Thanks
You can use the transaction command, https://docs.splunk.com/Documentation/Splunk/latest/SearchReference/Transaction
Assuming you have the field ID extracted, you can do
index=* | transaction ID
This will automatically produce a field called duration, which is the time between the first and last event with the same ID
While transaction will work, it's very inefficient
This stats should show you what you're looking for (presuming the fields are already extracted):
(index=ndxA OR index=ndxB) ID=* ("ADD" OR "REMOVE")
| stats min(_time) as when_added max(_time) as when_removed by ID
| eval when_added=strftime(when_added,"%c"), when_removed(when_removed,"%c")
If you don't already have fields extracted, you'll need to modify thusly (remove the "\D^" in the regex if the ID value isn't at the end of the line):
(index=ndxA OR index=ndxB) ("ADD" OR "REMOVE")
| rex field=_raw "ID \s+:\s+(?<ID>\d+)\D^"
| stats min(_time) as when_added max(_time) as when_removed by ID
| eval when_added=strftime(when_added,"%c"), when_removed(when_removed,"%c")
"daily unique entry text"
| spath input=stats
| where ('expectedCount' != _???_)
I have one daily unique log entry that states the expectedCount of items that will be processed for that day.
Let's say this daily entry contains the unique text daily unique entry text, and each time an item gets processed successfully, I log item processed.
I'd like an alert that fires if expectedCount is not equal to the number of item processed log entries that follow, in that day.
Can this be accomplished with something at _???_? Or: what's the best way to do this? Thanks in advance!
index=* "item processed" | stats count | append [ search index=* "daily unique entry text" | spath input=stats | fields expectedCount ] | stats values(count) as c, values(expectedCount) as ec | where c != ec
I think this is the most straightforward approach, but there are other ways. First search just gets a count of all the items processed. You may need to restrict it to a day of required. Then we append another search, which is just the value of the expected count. We use another stats command to get the value of the actual count and the expected count together. Then, just compare them