The Power of SPL
2 likes893 views
The document presents a comprehensive overview of Splunk's Search Processing Language (SPL), detailing its capabilities for data searching, filtering, and visualization, along with installation and setup procedures. It highlights the complexity and flexibility of SPL, including machine learning functionalities, various search command types, and performance optimization tips. Additionally, the document covers the structure of search pipelines, modes of searching, and practical examples for using SPL effectively.
![© 2017 SPLUNK INC.
▶ Enrich data with lookups
index=power_of_spl status!=200
| lookup customer_info uid
| stats count by customer_value
▶ Search Inception!
index=power_of_spl
[ search index=power_of_spl | stats sum(bytes)
as total_bytes by clientip
| sort - total_bytes | head 1 | return clientip ]
| stats count by clientip status uri | sort - count
▶ Append multiple searches
index=power_of_spl
| timechart span=15s avg(bytes) as avg_bytes
| appendcols [ search index=power_of_spl
| stats stdev(bytes) as stdev_bytes] | eval 2stdv_upper = avg_bytes + stdev_bytes*2 | filldown 2stdv_upper | eval 2stdv_lower = avg_bytes -
stdev_bytes*2 | filldown 2stdv_lower
| eval 2stdv_lower = if('2stdv_lower’ <0,0,'2stdv_lower') | fields - stdev_bytes
lookup – Converging Data Sources
Examples](https://image.slidesharecdn.com/ppt-splunkworkshop-2017-power-of-spl-170719145213/85/The-Power-of-SPL-47-320.jpg)
![© 2017 SPLUNK INC.
▶ Enrich data with lookups
index=power_of_spl status!=200
| lookup customer_info uid
| stats count by customer_value
▶ Search Inception!
index=power_of_spl
[ search index=power_of_spl | stats sum(bytes)
as total_bytes by clientip
| sort - total_bytes | head 1 | return clientip ]
| stats count by clientip status uri | sort - count
▶ Append multiple searches
index=power_of_spl
| timechart span=15s avg(bytes) as avg_bytes
| appendcols [ search index=power_of_spl
| stats stdev(bytes) as stdev_bytes] | eval 2stdv_upper = avg_bytes + stdev_bytes*2 | filldown 2stdv_upper | eval 2stdv_lower = avg_bytes -
stdev_bytes*2 | filldown 2stdv_lower
| eval 2stdv_lower = if('2stdv_lower’ <0,0,'2stdv_lower') | fields - stdev_bytes
Converging Data Sources
Examples](https://image.slidesharecdn.com/ppt-splunkworkshop-2017-power-of-spl-170719145213/85/The-Power-of-SPL-48-320.jpg)
![© 2017 SPLUNK INC.
▶ Enrich data with lookups
index=power_of_spl status!=200
| lookup customer_info uid
| stats count by customer_value
▶ Search Inception!
index=power_of_spl
[ search index=power_of_spl | stats sum(bytes)
as total_bytes by clientip
| sort - total_bytes | head 1 | return clientip ]
| stats count by clientip status uri | sort - count
▶ Append multiple searches
index=power_of_spl
| timechart span=15s avg(bytes) as avg_bytes
| appendcols [ search index=power_of_spl
| stats stdev(bytes) as stdev_bytes] | eval 2stdv_upper = avg_bytes + stdev_bytes*2 | filldown 2stdv_upper | eval 2stdv_lower =
avg_bytes - stdev_bytes*2 | filldown 2stdv_lower
| eval 2stdv_lower = if('2stdv_lower’ <0,0,'2stdv_lower') | fields - stdev_bytes
appendcols – Converging Data Sources
Examples](https://image.slidesharecdn.com/ppt-splunkworkshop-2017-power-of-spl-170719145213/85/The-Power-of-SPL-49-320.jpg)
![© 2017 SPLUNK INC.
Use Lookups
index=*
tag=authentication
action=failure user!=test
user!=test1 user!=test2
user!=test3 user!=test
user!=test4 user!=test5
user!=user
| stats count by user
| where count > 1000
| sort -count
index=*
tag=authentication
action=failure NOT
[|inputlookup
splunk_user_whitelist.csv
| fields user]
| stats count by user
| where count > 1000
| sort -count
Direct Exclusion Lookup Exclusion](https://image.slidesharecdn.com/ppt-splunkworkshop-2017-power-of-spl-170719145213/85/The-Power-of-SPL-104-320.jpg)
The Power of SPL
- 1. © 2017 SPLUNK INC.© 2017 SPLUNK INC. Power of Splunk Search Processing Language (SPL™) 07/13/2017 | Jacksonville Tolga Tohumcu | Staff Sales Engineer @TolgaTohumcu
- 2. © 2017 SPLUNK INC. During the course of this presentation, we may make forward-looking statements regarding future events or the expected performance of the company. We caution you that such statements reflect our current expectations and estimates based on factors currently known to us and that actual events or results could differ materially. For important factors that may cause actual results to differ from those contained in our forward-looking statements, please review our filings with the SEC. The forward-looking statements made in this presentation are being made as of the time and date of its live presentation. If reviewed after its live presentation, this presentation may not contain current or accurate information. We do not assume any obligation to update any forward-looking statements we may make. In addition, any information about our roadmap outlines our general product direction and is subject to change at any time without notice. It is for informational purposes only and shall not be incorporated into any contract or other commitment. Splunk undertakes no obligation either to develop the features or functionality described or to include any such feature or functionality in a future release. Splunk, Splunk>, Listen to Your Data, The Engine for Machine Data, Splunk Cloud, Splunk Light and SPL are trademarks and registered trademarks of Splunk Inc. in the United States and other countries. All other brand names, product names, or trademarks belong to their respective owners. © 2017 Splunk Inc. All rights reserved. Forward-Looking Statements
- 3. © 2017 SPLUNK INC. Set Up Before You Can Play Download the following at splunk.com and splunkbase.com ▶ Splunk Enterprise: • https://www.splunk.com/download ▶ Power of SPL App: • https://splunkbase.splunk.com/app/3353/
- 4. © 2017 SPLUNK INC. Set Up Before You Can Play ▶ Install Power of SPL App ▶ Now we’re ready to go!
- 5. © 2017 SPLUNK INC. ▶ License expired (already had older version installed) • Close browser, empty cache, open browser. If that doesn’t work: • Stop Splunk. • Uninstall all Splunk versions • Windows Control Panel->Uninstall programs->Splunk • OS X. Finder->Applications->Right click Splunk, Move to trash • Reinstall • Start Splunk ▶ Can’t start Splunk • Windows, Search Control panel ->Services->Splunk start • Linux; cd <SPLUNK dir>/splunk/bin;./splunk start Common problems at this point
- 6. © 2017 SPLUNK INC. 1. Installation and Setup (~15min) 2. Power of SPL Walkthrough (~1h 30min) • Overview & Anatomy of a Search • SPL Commands & Examples for Searching, charting, converging, mapping, transactions, anomalies, exploring data, custom 3. Custom Visualizations (~30min) 4. SPL and the Machine Learning Toolkit (~45min) Agenda
- 7. © 2017 SPLUNK INC. SPL Overview
- 8. © 2017 SPLUNK INC. ▶ Over 140 search commands ▶ Syntax was originally based upon the Unix pipeline and SQL and is optimized for time-series data ▶ The scope of SPL includes data searching, filtering, modification, manipulation, enrichment, insertion and deletion ▶ Includes machine learning such as anomaly detection SPL Overview Disk Intermediate results table Intermediate results table Final results table
- 9. © 2017 SPLUNK INC. ▶ Flexibility and effectiveness on small and big data ▶ Late-binding schema ▶ More/better methods of correlation ▶ Not just analyze, but visualize Why Create a New Query Language? Data BIG Data
- 10. © 2017 SPLUNK INC. search and filter | munge | report | cleanup | rename sum(KB) AS "Total KB" dc(clientip) AS "Unique Customers" | eval KB=bytes/1024 sourcetype=access* | stats sum(KB) dc(clientip) SPL Basic Structure
- 11. © 2017 SPLUNK INC. Knowledge Objects Data Interpretation – provide structure to raw data. (field extractions) Data Classification – group similar events. (event types, transactions) Data Enrichment – add value from external sources. (lookups) Data Normalization – group related fields. (tags, aliases) Data Models – represent one or more data sets. (pivot)
- 12. © 2017 SPLUNK INC. Types of Search Commands Streaming • Operate on each event individually • Distributable Streaming • Run on Indexers • Eval, fields, rename, regex • Centralized (Stateful) Streaming • Run on Search Head • head, streamstats 12 Transforming • Generate a reporting data structure • Operate on the entire event set • chart,timechart, stats, top Generating • Usually invoked at the beginning of a search • Do not expect or require an input • | search is implied • dbinspect, datamodel, inputcsv
- 13. © 2017 SPLUNK INC. Search Processor Categories Streaming (remote) • Consider each event/result row individually • Can be distributed to the indexers • Ex: eval, where, search, lookup Stateful Streaming • Consider each event/results row individually but maintain extra state modified by processing each event • Ordering is significant, and as such they force the search back to the search head • Ex: head, streamstats 13 Events • Need to see all events/result rows to perform their tasks • Ex: tail, sort, eventstats Transforming • Need to see all events/result rows and transform them into some new result rows • Stream Reporting • Can take batches of input, and then at any time generate output based on the input seen so far • Ex: stats, chart • Reporting • Must take all events at once • Ex: cluster, geostats
- 14. © 2017 SPLUNK INC. Search Pipelines Pipeline Setup ▶Search Pipelines are built according to the processors and the order they occur in the search string • At most, 5 pipelines are built, one for each processor category • Less restrictive processors may be accumulated into a more restrictive pipeline • Depending on the order of commands ▶Each processor is asked to first evaluate its arguments for validity, then processes them for setup 14
- 15. © 2017 SPLUNK INC. Search Pipelines Pipeline Processing ▶ Pipeline processing begins • Processors for the current category and earlier pipelines are accumulated in the current pipeline • Splunk moves to the next most restricted pipeline as we find a command that is restricted to that level • We never move back to a less restrictive pipeline once moving to a more restrictive pipeline ▶ Streaming Pipeline • Only pipeline that can be distributed to indexers • After all streaming commands are accumulated, non-streaming commands are asked if they can be split • Ex: Prestats, Presort, etc… • If available, commands are added to the Streaming Pipeline 15
- 16. © 2017 SPLUNK INC. Search Pipelines Pipeline Rendering ▶Streaming Pipeline • Rendered as remoteSearch string and sent to Indexers ▶Stateful and Events Pipelines • Rendered into eventsSearch ▶Stream Report and Report Pipelines • Rendered into reportSearch 16
- 17. © 2017 SPLUNK INC. Search Modes ● Historical – Cursored Search ‣ Most historical searches run in this mode ‣ Returns events in strict time ordering – Batch Mode Search ‣ Automatically enabled when possible ‣ Reads buckets sequentially = returns data faster ‣ Events are not time ordered ‣ Uses more memory ‣ Job Inspector: isBatchModeSearch 17 ● Realtime – Realtime Search ‣ Impacts indexing performance ‣ Hooks into indexing pipeline daemon to build a separate queue of events – Indexed Realtime Search ‣ Continuous search ‣ Does not hook into indexing pipeline ‣ Default latency of 60sec with 1 sec lookback – Parameters can be adjusted
- 18. © 2017 SPLUNK INC. Search UI ● Search UI Modes – Fast, Smart, Verbose Mode – Determine which fields are extracted – Turn timeliner on/off ● Data Preview – When its time, main search thread is stopped so Splunk can gather current results and generate a preview – Lots of parameters available to tweak this – If you don’t need a preview, disable it in the UI ● Timeliner – Runs in the background – May cause unnecessary events to be downloaded, especially when in verbose mode – By default, 1,000 events are downloaded per column with 300 columns max – Parameters available to adjust this or turn it off 18
- 19. © 2017 SPLUNK INC. Search Advice ● Filter early – Specify an index – Utilize indexed extractions where available – Use the TERM directive if applicable ● Place streaming/remote commands before non-streaming commands ● Avoid using table, except at the very end – This is a reporting command and will cause data to be pushed to Search Head ● Remove unnecessary data using | fields 19
- 20. © 2017 SPLUNK INC. Job Inspector Types of Components ▶ Command • Measures actual work being done ▶ Dispatch • Measures Splunk framework 20 ● Invocations – # of times the component was run ● Input/Output Count – Generally the event count – Exception: dispatch.stream.remote ‣ Byte count of data received from the peer
- 21. © 2017 SPLUNK INC. Job Inspector Dispatch.evaluate ▶ Search is parsed here ▶ Sub-searches are run here 21
- 22. © 2017 SPLUNK INC. Job Inspector Dispatch.stream.remote ▶ Time spent searching on peers ▶ Useful for comparing time spent and # of results returned across peers 22
- 23. © 2017 SPLUNK INC. Job Inspector Command.search.index.usec ▶ Indicates how long it took to search an index in microseconds • Helps identify slow searching of indexes • Can indicate VERY large indexes (especially on rare term searches) • Lots of indexed extractions? Data with extremely high cardinality? • Can indicate slow disk or disk contention (especially on dense searches) ▶ Histogram of invocations grouped by microseconds ▶ Correlate this to dispatch.stream.remote 23
- 24. © 2017 SPLUNK INC. Job Inspector Other Interesting Components… ▶ Command.search.rawdata • Time peer spend uncompressing bucket slices • On a dense search, can indicate CPU contention ▶ Command.search.filter • Post filtering after KV extraction (schema-on-the-fly) ▶ Command.search.createProviderQueue • Time spent setting up connections to peers 24
- 25. © 2017 SPLUNK INC. SPL Examples
- 26. © 2017 SPLUNK INC. ▶ Find the needle in the haystack ▶ Charting statistics and predicting values ▶ Enriching and converging data sources ▶ Map geographic data in real time (We’ll add some custom viz here!) ▶ Identifying anomalies ▶ Transactions ▶ Data exploration & finding relationships between fields ▶ Custom commands SPL Examples and Recipes
- 27. © 2017 SPLUNK INC. ▶ Find the needle in the haystack ▶ Charting statistics and predicting values ▶ Enriching and converging data sources ▶ Map geographic data in real time ▶ Identifying anomalies ▶ Transactions ▶ Data exploration & finding relationships between fields ▶ Custom commands SPL Examples and Recipes
- 28. © 2017 SPLUNK INC. ▶ Keyword search: sourcetype=access* http ▶ Filter: sourcetype=access* http host=webserver-02 ▶ Combined: sourcetype=access* http host=webserver-02 (503 OR 504) Search and Filter Examples
- 29. © 2017 SPLUNK INC. ▶ Keyword search: sourcetype=access* http ▶ Filter: sourcetype=access* http host=webserver-02 ▶ Combined: sourcetype=access* http host=webserver-02 (503 OR 504) Search and Filter Examples
- 30. © 2017 SPLUNK INC. ▶ Keyword search: sourcetype=access* http ▶ Filter: sourcetype=access* http host=webserver-02 ▶ Combined: sourcetype=access* http host=webserver-02 (503 OR 504) Search and Filter Examples
- 31. © 2017 SPLUNK INC. ▶ Calculation: sourcetype=access* |eval KB=bytes/1024 ▶ Evaluation: sourcetype=access* | eval http_response = if(status != 200, ”Error", ”OK”) ▶ Concatenation: sourcetype=access* | eval connection = device.” - ".clientip Eval – Modify or Create New Fields and Values Examples
- 32. © 2017 SPLUNK INC. ▶ Calculation: sourcetype=access* |eval KB=bytes/1024 ▶ Evaluation: sourcetype=access* | eval http_response = if(status != 200, ”Error", ”OK”) ▶ Concatenation: sourcetype=access* | eval connection = device.” - ".clientip Eval – Modify or Create New Fields and Values Examples
- 33. © 2017 SPLUNK INC. ▶ Calculation: sourcetype=access* |eval KB=bytes/1024 ▶ Evaluation: sourcetype=access* | eval http_response = if(status != 200, ”Error", ”OK”) ▶ Concatenation: sourcetype=access* | eval connection = device.” - ".clientip Eval – Modify or Create New Fields and Values Examples
- 34. © 2017 SPLUNK INC. Eval – Just Getting Started! Splunk Search Quick Reference Guide
- 35. © 2017 SPLUNK INC. ▶ Find the needle in the haystack ▶ Charting statistics and predicting values ▶ Enriching and converging data sources ▶ Map geographic data in real time ▶ Identifying anomalies ▶ Transactions ▶ Data exploration & finding relationships between fields ▶ Custom commands SPL Examples and Recipes
- 36. © 2017 SPLUNK INC. ▶ Calculate stats and rename Index=power_of_spl | stats avg(bytes) AS “Avg Bytes” ▶ Multiple statistics index=power_of_spl | stats avg(bytes) AS bytes sparkline(avg(bytes)) AS Bytes_Trend min(bytes) max(bytes) ▶ By another field index=power_of_spl | stats avg(bytes) AS avg_bytes sparkline(avg(bytes)) AS Bytes_Trend min(bytes) max(bytes) by clientip | sort - avg_bytes Stats – Calculate Statistics Based on Field Values Examples
- 37. © 2017 SPLUNK INC. ▶ Calculate stats and rename Index=power_of_spl | stats avg(bytes) AS “Avg Bytes” ▶ Multiple statistics index=power_of_spl | stats avg(bytes) AS bytes sparkline(avg(bytes)) AS Bytes_Trend min(bytes) max(bytes) ▶ By another field index=power_of_spl | stats avg(bytes) AS avg_bytes sparkline(avg(bytes)) AS Bytes_Trend min(bytes) max(bytes) by clientip | sort - avg_bytes Stats – Calculate Statistics Based on Field Values Examples
- 38. © 2017 SPLUNK INC. ▶ Calculate stats and rename Index=power_of_spl | stats avg(bytes) AS “Avg Bytes” ▶ Multiple statistics index=power_of_spl | stats avg(bytes) AS bytes sparkline(avg(bytes)) AS Bytes_Trend min(bytes) max(bytes) ▶ By another field index=power_of_spl | stats avg(bytes) AS avg_bytes sparkline(avg(bytes)) AS Bytes_Trend min(bytes) max(bytes) by clientip | sort - avg_bytes Stats – Calculate Statistics Based on Field Values Examples
- 39. © 2017 SPLUNK INC. ▶ Visualize stats over time index=power_of_spl | timechart avg(bytes) ▶ Add a trendline index=power_of_spl | timechart avg(bytes) as bytes | trendline sma5(bytes) ▶ Add a prediction overlay index=power_of_spl | timechart avg(bytes) as bytes | predict future_timespan=5 bytes Timechart – Visualize Statistics Over Time Examples
- 40. © 2017 SPLUNK INC. ▶ Visualize stats over time index=power_of_spl | timechart avg(bytes) ▶ Add a trendline index=power_of_spl | timechart avg(bytes) as bytes | trendline sma5(bytes) ▶ Add a prediction overlay index=power_of_spl | timechart avg(bytes) as bytes | predict future_timespan=5 bytes Timechart – Visualize Statistics Over Time Examples
- 41. © 2017 SPLUNK INC. ▶ Visualize stats over time index=power_of_spl | timechart avg(bytes) ▶ Add a trendline index=power_of_spl | timechart avg(bytes) as bytes | trendline sma5(bytes) ▶ Add a prediction overlay index=power_of_spl | timechart avg(bytes) as bytes | predict future_timespan=5 bytes Timechart – Visualize Statistics Over Time Examples
- 42. © 2017 SPLUNK INC. ▶ Cumulative/Running Totals index=power_of_spl | reverse | streamstats sum(bytes) AS sum_bytes | timechart latest(sum_bytes) as "Total Bytes" ▶ Summary Statistics index=power_of_spl | eventstats avg(bytes) AS overall_avg_bytes | stats avg(bytes) as clientip_avg_bytes by clientip overall_avg_bytes Streamstats – Cumulative/Running Totals Statistics Examples
- 43. © 2017 SPLUNK INC. ▶ Cumulative/Running Totals index=power_of_spl | reverse | streamstats sum(bytes) AS sum_bytes | timechart latest(sum_bytes) as "Total Bytes" ▶ Summary Statistics index=power_of_spl | eventstats avg(bytes) AS overall_avg_bytes | stats avg(bytes) as clientip_avg_bytes by clientip overall_avg_bytes Streamstats – Cumulative/Running Totals Statistics Examples
- 44. © 2017 SPLUNK INC. Stats/Timechart – But Wait, There’s More! Splunk Search Quick Reference Guide
- 45. © 2017 SPLUNK INC. ▶ Find the needle in the haystack ▶ Charting statistics and predicting values ▶ Enriching and converging data sources ▶ Map geographic data in real time ▶ Identifying anomalies ▶ Transactions ▶ Data exploration & finding relationships between fields ▶ Custom commands SPL Examples and Recipes
- 46. © 2017 SPLUNK INC. Converging Data Sources Index Untapped Data: Any Source, Type, Volume Ask Any Question Application Delivery Security, Compliance and Fraud IT Operations Business Analytics Industrial Data and the Internet of Things On-Premises Private Cloud Public Cloud Storage Online Shopping Cart Telecoms Desktops Security Web Services Networks Containers Web Clickstreams RFID Smartphones and Devices Servers Messaging GPS Location Packaged Applications Custom Applications Online Services DatabasesCall Detail Records Energy Meters Firewall Intrusion Prevention
- 47. © 2017 SPLUNK INC. ▶ Enrich data with lookups index=power_of_spl status!=200 | lookup customer_info uid | stats count by customer_value ▶ Search Inception! index=power_of_spl [ search index=power_of_spl | stats sum(bytes) as total_bytes by clientip | sort - total_bytes | head 1 | return clientip ] | stats count by clientip status uri | sort - count ▶ Append multiple searches index=power_of_spl | timechart span=15s avg(bytes) as avg_bytes | appendcols [ search index=power_of_spl | stats stdev(bytes) as stdev_bytes] | eval 2stdv_upper = avg_bytes + stdev_bytes*2 | filldown 2stdv_upper | eval 2stdv_lower = avg_bytes - stdev_bytes*2 | filldown 2stdv_lower | eval 2stdv_lower = if('2stdv_lower’ <0,0,'2stdv_lower') | fields - stdev_bytes lookup – Converging Data Sources Examples
- 48. © 2017 SPLUNK INC. ▶ Enrich data with lookups index=power_of_spl status!=200 | lookup customer_info uid | stats count by customer_value ▶ Search Inception! index=power_of_spl [ search index=power_of_spl | stats sum(bytes) as total_bytes by clientip | sort - total_bytes | head 1 | return clientip ] | stats count by clientip status uri | sort - count ▶ Append multiple searches index=power_of_spl | timechart span=15s avg(bytes) as avg_bytes | appendcols [ search index=power_of_spl | stats stdev(bytes) as stdev_bytes] | eval 2stdv_upper = avg_bytes + stdev_bytes*2 | filldown 2stdv_upper | eval 2stdv_lower = avg_bytes - stdev_bytes*2 | filldown 2stdv_lower | eval 2stdv_lower = if('2stdv_lower’ <0,0,'2stdv_lower') | fields - stdev_bytes Converging Data Sources Examples
- 49. © 2017 SPLUNK INC. ▶ Enrich data with lookups index=power_of_spl status!=200 | lookup customer_info uid | stats count by customer_value ▶ Search Inception! index=power_of_spl [ search index=power_of_spl | stats sum(bytes) as total_bytes by clientip | sort - total_bytes | head 1 | return clientip ] | stats count by clientip status uri | sort - count ▶ Append multiple searches index=power_of_spl | timechart span=15s avg(bytes) as avg_bytes | appendcols [ search index=power_of_spl | stats stdev(bytes) as stdev_bytes] | eval 2stdv_upper = avg_bytes + stdev_bytes*2 | filldown 2stdv_upper | eval 2stdv_lower = avg_bytes - stdev_bytes*2 | filldown 2stdv_lower | eval 2stdv_lower = if('2stdv_lower’ <0,0,'2stdv_lower') | fields - stdev_bytes appendcols – Converging Data Sources Examples
- 50. © 2017 SPLUNK INC. ▶ Find the needle in the haystack ▶ Charting statistics and predicting values ▶ Enriching and converging data sources ▶ Map geographic data in real time (Let’s add some viz!) ▶ Identifying anomalies ▶ Transactions ▶ Data exploration & finding relationships between fields ▶ Custom commands SPL Examples and Recipes
- 51. © 2017 SPLUNK INC. ▶ Assign Lat/Lon to IP addresses … | iplocation clientip ▶ Visualize statistics geographically … | geostats sum(price) by product ▶ Use custom choropleths … | geom <featureCollection> <featureId> ▶ Track object movements … | table _time latitude longitude vehicleId iplocation Geographic Data Examples
- 52. © 2017 SPLUNK INC. ▶ Assign Lat/Lon to IP addresses … | iplocation clientip ▶ Visualize statistics geographically … | geostats sum(price) by product ▶ Use custom choropleths … | geom <featureCollection> <featureId> ▶ Track object movements … | table _time latitude longitude vehicleId geostats – Geographic Data Examples
- 53. © 2017 SPLUNK INC. ▶ Assign Lat/Lon to IP addresses … | iplocation clientip ▶ Visualize statistics geographically … | geostats sum(price) by product ▶ Use custom choropleths … | geom <featureCollection> <featureId> ▶ Track object movements … | table _time latitude longitude vehicleId geom – Geographic Data Examples
- 54. © 2017 SPLUNK INC. ▶ Assign Lat/Lon to IP addresses … | iplocation clientip ▶ Visualize statistics geographically … | geostats sum(price) by product ▶ Use custom choropleths … | geom <featureCollection> <featureId> ▶ Track object movements … | table _time latitude longitude vehicleId table – Geographic Data Examples
- 55. © 2017 SPLUNK INC. Custom Visualizations
- 56. © 2017 SPLUNK INC. ▶ Native charts and maps • Bar / Line / Area charts • Bubble / Scatter plots • Gauges • Maps • Single Value Displays • Tables ▶ Generalized to fit use cases across many different areas ▶ Can be customized to some extent to cover specific use cases Native Visualizations In Splunk 56
- 57. © 2017 SPLUNK INC. ▶ Many use cases require a more specific visualization ▶ Specific custom appearance ▶ Represent data where native visualizations are not suitable • You can Splunk everything! • We won’t be able to predict every possible use case • Still uses SPL to drive visualizations Custom Visualizations FTW!
- 58. © 2017 SPLUNK INC. ▶ Platform extensibility framework and API ▶ Targeted at internal and external developers with web development / JS skills and basic knowledge of the Splunk platform ▶ Developers can make use of any third party libraries (d3.js, three.js, highcharts.js, etc…) that run in the browser* * with minor adjustments, and if third party license permits such use Custom Visualizations
- 59. © 2017 SPLUNK INC. Custom Visualizations For Admins In-product • Packaged as an app! • Installed like any other app • Users can search for visualizations on Splunkbase and directly in the product Installation
- 60. © 2017 SPLUNK INC. ▶ Choose from potentially dozens of installed visualizations! ▶ Appears as a first-class citizen alongside native visualizations • Looks and works just like packaged native visualizations ▶ Customize functionality and appearance of the visualization without touching any code, straight from the UI ▶ SPL Example provided as you hover over each visualization option. Custom Visualizations How-to
- 61. © 2017 SPLUNK INC. New Splunk Visualizations Treemap Sankey Diagram Punchcard Calendar Heat Map Parallel Coordinates Bullet GraphLocation Tracker Horseshoe Meter Machine Learning Charts Timeline Horizon Chart Multiple use cases across IT, security, IoT, and business analytics
- 62. © 2017 SPLUNK INC. Box Plot 3D scatter plot New Partner/Community Visualizations Wordcloud Donut Chart Heat Map
- 63. © 2017 SPLUNK INC. New Partner/Community Visualizations Geo Heatmap Custom Cluster Map Clustered Single Value Map Missile Map
- 64. © 2017 SPLUNK INC. Custom Visualizations – Demo!
- 65. © 2017 SPLUNK INC. Demo Screenshot #1
- 66. © 2017 SPLUNK INC. Demo Screenshot #2
- 67. © 2017 SPLUNK INC. ▶ Find the needle in the haystack ▶ Charting statistics and predicting values ▶ Enriching and converging data sources ▶ Map geographic data in real time ▶ Identifying anomalies ▶ Transactions ▶ Data exploration & finding relationships between fields ▶ Custom commands SPL Examples and Recipes
- 68. © 2017 SPLUNK INC. ▶ Find anomalies | inputlookup car_data.csv | anomalydetection ▶ Summarize anomalies | inputlookup car_data.csv | anomalydetection action=summary ▶ Use IQR and remove outliers | inputlookup car_data.csv | anomalydetection method=iqr action=remove Anomaly Detection – Find anomalies in your data Examples
- 69. © 2017 SPLUNK INC. ▶ Find the needle in the haystack ▶ Charting statistics and predicting values ▶ Enriching and converging data sources ▶ Map geographic data in real time ▶ Identifying anomalies ▶ Transactions ▶ Data exploration & finding relationships between fields ▶ Custom commands SPL Examples and Recipes
- 70. © 2017 SPLUNK INC. ▶ Group by session ID sourcetype=access* | transaction JSESSIONID ▶ Calculate session durations sourcetype=access* | transaction JSESSIONID | stats min(duration) max(duration) avg(duration) ▶ Stats is better sourcetype=access* | stats min(_time) AS earliest max(_time) AS latest by JSESSIONID | eval duration=latest-earliest | stats min(duration) max(duration) avg(duration) Transaction – Group Related Events Spanning Time Examples
- 71. © 2017 SPLUNK INC. ▶ Group by session ID sourcetype=access* | transaction JSESSIONID ▶ Calculate session durations sourcetype=access* | transaction JSESSIONID | stats min(duration) max(duration) avg(duration) ▶ Stats is better sourcetype=access* | stats min(_time) AS earliest max(_time) AS latest by JSESSIONID | eval duration=latest-earliest | stats min(duration) max(duration) avg(duration) Transaction – Group Related Events Spanning Time Examples
- 72. © 2017 SPLUNK INC. ▶ Group by session ID sourcetype=access* | transaction JSESSIONID ▶ Calculate session durations sourcetype=access* | transaction JSESSIONID | stats min(duration) max(duration) avg(duration) ▶ Stats is better sourcetype=access* | stats min(_time) AS earliest max(_time) AS latest by JSESSIONID | eval duration=latest-earliest | stats min(duration) max(duration) avg(duration) Transaction – Group Related Events Spanning Time Examples
- 73. © 2017 SPLUNK INC. ▶ Find the needle in the haystack ▶ Charting statistics and predicting values ▶ Enriching and converging data sources ▶ Map geographic data in real time ▶ Identifying anomalies ▶ Transactions ▶ Data exploration & finding relationships between fields ▶ Custom commands SPL Examples and Recipes
- 74. © 2017 SPLUNK INC. Data Exploration | analyzefields | anomalies | arules | associate | cluster | contingency | correlate | fieldsummary
- 75. © 2017 SPLUNK INC. ▶ Find most/least common events * | cluster showcount=t t=.1 | table _raw cluster_count ▶ Display Summary of Fields sourcetype=access_combined | fields – date* source* time* | fieldsummary maxvals=5 ▶ Show patterns of co-occurring fields sourcetype=access_combined | fields – date* source* time* | correlate ▶ View field relationships sourcetype=access_combined | contingency uri status ▶ Find predictors of fields sourcetype=access_combined | analyzefields classfield=status Cluster – Exploring Your Data (find common and/or rare events within your data) Examples
- 76. © 2017 SPLUNK INC. ▶ Find most/least common events * | cluster showcount=t t=.1 | table _raw cluster_count ▶ Display Summary of Fields sourcetype=access_combined | fields – date* source* time* | fieldsummary maxvals=5 ▶ Show patterns of co-occurring fields sourcetype=access_combined | fields – date* source* time* | correlate ▶ View field relationships sourcetype=access_combined | contingency uri status ▶ Find predictors of fields sourcetype=access_combined | analyzefields classfield=status fieldsummary – Exploring Your Data Gives you a quick breakdown of your numerical fields such as count, min, max, stdev, etc. It also shows you examples values in the event. Examples
- 77. © 2017 SPLUNK INC. ▶ Find most/least common events * | cluster showcount=t t=.1 | table _raw cluster_count ▶ Display Summary of Fields sourcetype=access_combined | fields – date* source* time* | fieldsummary maxvals=5 ▶ Show patterns of co-occurring fields sourcetype=access_combined | fields – date* source* time* | correlate ▶ View field relationships sourcetype=access_combined | contingency uri status ▶ Find predictors of fields sourcetype=access_combined | analyzefields classfield=status correlate – Exploring Your Data find co-occurrence between fields. Basically a matrix showing the ‘Field1 exists 80% of the time when Field2 exists Examples
- 78. © 2017 SPLUNK INC. ▶ Find most/least common events * | cluster showcount=t t=.1 | table _raw cluster_count ▶ Display Summary of Fields sourcetype=access_combined | fields – date* source* time* | fieldsummary maxvals=5 ▶ Show patterns of co-occurring fields sourcetype=access_combined | fields – date* source* time* | correlate ▶ View field relationships sourcetype=access_combined | contingency uri status ▶ Find predictors of fields sourcetype=access_combined | analyzefields classfield=status contingency – Exploring Your Data look for relationships of between two fields. Examples
- 79. © 2017 SPLUNK INC. ▶ Find most/least common events * | cluster showcount=t t=.1 | table _raw cluster_count ▶ Display Summary of Fields sourcetype=access_combined | fields – date* source* time* | fieldsummary maxvals=5 ▶ Show patterns of co-occurring fields sourcetype=access_combined | fields – date* source* time* | correlate ▶ View field relationships sourcetype=access_combined | contingency uri status ▶ Find predictors of fields sourcetype=access_combined | analyzefields classfield=status analyzefields – Exploring Your Data extremely useful for not only looking for meaningful fields in your data, but also for determining which fields to use in linear or logistical regression algorithms in the machine learning appExamples
- 80. © 2017 SPLUNK INC. ▶ Predict Numeric Fields ▶ Predict Categorical Fields ▶ Detect Numerical Outliers ▶ Detect Categorical Outliers ▶ Forecast Time Series ▶ Cluster Events Machine Learning Toolkit and Showcase Examples
- 81. © 2017 SPLUNK INC. ▶ Find the needle in the haystack ▶ Charting statistics and predicting values ▶ Enriching and converging data sources ▶ Map geographic data in real time ▶ Identifying anomalies ▶ Transactions ▶ Data exploration & finding relationships between fields ▶ Custom commands SPL Examples and Recipes
- 82. © 2017 SPLUNK INC. ▶ What is a Custom Command? • “| haversine origin="47.62,-122.34" outputField=dist lat lon” ▶ Why do we use Custom Commands? • Run other/external algorithms on your Splunk data • Save time munging data (see Timewrap!) • Because you can! ▶ Create your own or download as Apps • Haversine (Distance between two GPS coords) • Timewrap (Enhanced Time overlay) • Levenshtein (Fuzzy string compare) • Base64 (Encode/Decode) Custom Commands
- 83. © 2017 SPLUNK INC. ▶ Download and install App Haversine ▶ Read documentation then use in SPL! sourcetype=access* | iplocation clientip | search City=A* | haversine origin="47.62,-122.34" units=mi outputField=dist lat lon | table clientip, City, dist, lat, lon Custom Commands – Haversine Examples
- 84. © 2017 SPLUNK INC. ▶ Download and install App Haversine ▶ Read documentation then use in SPL! sourcetype=access* | iplocation clientip | search City=A* | haversine origin="47.62,-122.34" units=mi outputField=dist lat lon | table clientip, City, dist, lat, lon Custom Commands – Haversine Examples
- 85. © 2017 SPLUNK INC. SPL & The Machine Learning Toolkit
- 86. © 2017 SPLUNK INC. ▶ Predict Numeric Fields ▶ Predict Categorical Fields ▶ Detect Numerical Outliers ▶ Detect Categorical Outliers ▶ Forecast Time Series ▶ Cluster Events Machine Learning Toolkit and Showcase Examples
- 87. © 2017 SPLUNK INC. Machine Learning with the Splunk Platform Visualize Share Clean Transform Operationalize Monitor Alert Build Model Search Explore Collect Data Test, Improve Models Ecosystem MLTK Choose Algorithm Ecosystem Splunk Splunk Splunk Splunk MLTK Splunk MLTK Splunk MLTK Splunk Ecosystem Splunk Real-time Data Science Pipeline Ecosystem MLTK Splunk Splunk’s App Ecosystem contains 1000’s of free add-ons for getting data in, applying structure and visualizing your data giving you faster time to value. The Machine Learning Toolkit delivers new SPL commands, custom visualizations, assistants, and examples to explore a variety of ml concepts. Splunk Enterprise is the mission-critical platform for indexing, searching, analyzing, alerting and visualizing machine data. Packaged: UBA, ITSI
- 88. © 2017 SPLUNK INC. ML SPL Visualize Share Correlate Clean Munge Operationalize Monitor Alert Build Model Search Explore Universal Indexing Test, Improve Models Ecosystem MLTK Choose Algorithm Ecosystem Splunk Splunk Splunk Splunk MLTK Splunk MLTK Splunk MLTK Splunk Ecosystem Splunk fit sample apply listmodels deletemodel summary eval rex stats eventstats streamstats table … timechart chart stats geostats geom sendalert sendemail table … MLTK Library predict (cmd) anomalydetection (cmd) analyzefields anomalies arules associate cluster contingency correlate fieldsummary …
- 89. © 2017 SPLUNK INC. MLTK Commands The Machine Learning Toolkit contains several custom search commands that implement classic machine learning and statistical learning tasks: • fit: Fit and apply a machine learning model to search results. • apply: Apply a machine learning model that was learned using the fit command. • summary: Return a summary of a machine learning model that was learned using the fit command. • listmodels: Return a list of machine learning models that were learned using the fit command. • deletemodel: Delete a machine learning model that was learned using the fit command. • sample: Randomly sample or partition events.
- 90. © 2017 SPLUNK INC. ML-SPL Demo
- 91. © 2017 SPLUNK INC. Set Up Before You Can Play Download the following at splunkbase.com ▶ Machine Learning Toolkit: • https://splunkbase.splunk.com/app/2890/ ▶ Python for Scientific Computing: • https://splunkbase.splunk.com/app/2881/ *Note – For the Python for Scientific Computing App you need to download the platform specific version – Mac, Linux, Windows
- 92. © 2017 SPLUNK INC. | fit
- 93. © 2017 SPLUNK INC. | sample
- 94. © 2017 SPLUNK INC. | apply
- 95. © 2017 SPLUNK INC. | listmodels
- 96. © 2017 SPLUNK INC. | summary
- 97. © 2017 SPLUNK INC. | deletemodel
- 98. © 2017 SPLUNK INC.© 2017 SPLUNK INC. Best Practices
- 99. © 2017 SPLUNK INC. The Truth about SPL COMMAND ORDER MATTERS
- 100. © 2017 SPLUNK INC. Command Order Streaming > Transforming
- 101. © 2017 SPLUNK INC. LoadJob • Loads events or results of a previously completed search job. • Identified either by the search job id or a scheduled search name. One Search, Many Panels | loadjob savedsearch=”Name of Saved Search"
- 102. © 2017 SPLUNK INC. One Search, Many Panels | loadjob savedsearch=”Name of Saved Search" | eval threat_ip=if(match_field="src_ip",src_ip,dest_ip) | eval start_time=coalesce(relative_time(now(),"$field1.earliest$"),"- 30d@m") | eval end_time=coalesce(relative_time(now(),"$field1.latest$"),now()) | where _time>=start_time AND _time<=end_time | timechart span=1h dc(threat_ip) as threat_ip_count by feed_category
- 103. © 2017 SPLUNK INC. One Search, Many Panels Macros • Reusable chunks of SPL that you can use within other searches. • Allows for the passing of arguments. (ex. |`splunk_domain_split(url)`)
- 104. © 2017 SPLUNK INC. Use Lookups index=* tag=authentication action=failure user!=test user!=test1 user!=test2 user!=test3 user!=test user!=test4 user!=test5 user!=user | stats count by user | where count > 1000 | sort -count index=* tag=authentication action=failure NOT [|inputlookup splunk_user_whitelist.csv | fields user] | stats count by user | where count > 1000 | sort -count Direct Exclusion Lookup Exclusion
- 105. © 2017 SPLUNK INC. ▶Additional information can be found in: • Power of SPL App! • Docs - Search Manual • Docs - MLTK Search Commands • MLTK Quick Reference Guide • Blogs • Answers • Exploring Splunk For More Information
- 106. © 2017 SPLUNK INC. • SPL Examples App Other Useful Apps to download! • Splunk 6.x Dashboard Examples • Splunk 6.5 Overview App
- 107. © 2017 SPLUNK INC. ▶ Benefit of using your own laptop is a ‘Take it with you after’ approach – and bang on it without messing up production Splunk • Promote Personalized 50GB 6 month Dev/Test license to use for Workshop (vs. 30 day download trial) • Encourages long term playing with Splunk and continuously testing out new data sources How about an extra 50GB?
- 108. © 2017 SPLUNK INC. Q & A
- 109. © 2017 SPLUNK INC.© 2017 SPLUNK INC. Thank You