JohnTheLearner opened a new issue #4676: Temporal overlap of anomalous events
URL: https://github.com/apache/incubator-pinot/issues/4676
 
 
   I noticed anomalous event overlaps while running ThirdEye over the demo 
data.  Configuration parameters are available to specify if, and how large, 
gaps between anomalous events are allowed to be, but anomalous events should be 
exclusive of each other (no temporal overlap).
   
   ### Overlap (zoomed)
   In the below example, the 8/18 and 8/23 events overlap by ~ 2 days:
   ![TE overlap - zoomed 
out](https://user-images.githubusercontent.com/9376440/66166379-66bdb380-e5f4-11e9-9e02-98d877df2eda.png)
   
   ### Time Range
   Since the demo data is generated each time it’s run, I doubt the time range 
is significant, but here is the range I observed during this case (this wasn't 
the only observation of this, though):
   ![TE overlap - time 
range](https://user-images.githubusercontent.com/9376440/66166420-8359eb80-e5f4-11e9-9170-f4d8afd8058f.png)
   
   ### Entire Preview 
   ...[when setting-up the 
detection)](http://localhost:51426/app/#/self-serve/create-alert) was:  
   ![TE overlap - zoomed 
out](https://user-images.githubusercontent.com/9376440/66166543-d6cc3980-e5f4-11e9-96e0-50f94d2cca35.png)
   
   ### Detection Configuration 
   ...that generated the above output:
   ```
   # Below is a sample template. You may refer the documentation for more 
examples and update the fields accordingly.
   
   # Give a name for this anomaly detection pipeline (should be unique).
   detectionName: detection_test_1
   
   # Tell the alert recipients what it means if this alert is fired.
   description: If this alert fires then it means so-and-so and check so-and-so 
for irregularities
   
   # The metric you want to do anomaly detection on. You may type a few 
characters and look ahead (ctrl + space) to auto-fill.
   metric: purchases
   
   # The dataset or UMP table name to which the metric belongs. Look ahead 
should auto populate this field.
   dataset: business
   
   rules:
   - detection:
      - name: detection_rule_1
        type: PERCENTAGE_RULE
        params:
          offset: do1d
          percentageChange: 0.01
   
   - detection:
      - name: detection_rule_2
        type: THRESHOLD
        params:
          min: 140
   ```
   

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
[email protected]


With regards,
Apache Git Services

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to