DBMS Final Project

Ref

Anomaly Detection with InfluxDB Tutorial This tutorial is poorly organized, we provide more thorough instructions here.

Install Dependencies

• python3
• InfluxDB==1.7.9
• Kapacitor==1.5.3 https://portal.influxdata.com/downloads/

Start Services

Mac

brew services start influxdb
brew services start kapacitor
brew services start telegraf

Ubuntu

sudo systemctl start influxdb
sudo systemctl start kapacitor
sudo systemctl start telegraf

Verify Services Status

Execute the command and check for errors.

Mac

TODO

Ubuntu

sudo journalctl -f -n 128 -u influxdb

or

sudo systemctl status <influxdb/kapacitor/telegraf>

Instructions

You can run bash run.sh directly, detailed descriptions are below.

1. Change Kapacitor configs Append the code in kapacitor-part.conf to your kapacitor configuration file. It's typically located at /etc/kapacitor/kapacitor.conf Note that you should remove the previously defined [udf] function if there is any in your conf file.

2. Restart Kapacitor

   sudo service kapacitor restart

   Check logs for errors

   sudo tail -f -n 128 /var/log/kapacitor/kapacitor.log

3. Define Kapacitor Task

   kapacitor define print_temps -tick print_temps.tick

   If success, there should be nothing printed to your console.

4. Start the recording in the background

   rid=$(kapacitor record stream -task print_temps -duration 24h -no-wait)

5. Run python script to generate data

   python printer_data.py

6. We can verify it worked by listing information about the recording. Our recording came out to 1.6MB, so yours should come out somewhere close to that:

   $ kapacitor list recordings $rid
   ID                                      Type    Status    Size      Date
   7bd3ced5-5e95-4a67-a0e1-f00860b1af47    stream  finished  1.6 MB    04 May 16 11:44 MDT
   

7. Finally, let’s run the play against our task and see how it works:

   kapacitor replay -task print_temps -recording $rid -rec-time

8. Check the various log files to see if the algorithm caught the anomalies:

   cat /tmp/kapacitor_udf/{hotend,bed,air}_failure.log