Redshift Performance Problem

Options
Figauro
Figauro Member
in Magic ETL

I have call center data that I pull in through inContact and I need to add a date field from a date/time field in order to not have problems with time. To do this, I've made a redshift flow that selects all the fields and truncates the date/time field to a date field:

 

Select field1, field2, ...,

TRUNC(CONVERT_TIMEZONE('UTC','America/Chicago',"contactStart")) AS contactStartdate

from table

 

Since the data comes in as UTC, I first convert it to central and the truncate it. That way I get the correct date out of the stamp. The table has 52 columns and usually about 3000-4000 rows being replaced every 15min.

 

My problem is I need to update this dataset every 15min, but the redshift query sometime takes 30 seconds or over 10-15min. Any ideas on what is going on/how I can get it to consistently run for only 30 seconds?

 

Best Answer

  • DDalt
    DDalt Member
    Answer ✓
    Options

    I was curious about Redshift run times being inconsistent and posed a question on DOJO day. Here is the question and response:

     

    https://dojo.domo.com/t5/Beast-Mode-ETL-Dataflow/Redshift-vs-MySQL-vs-ETL/m-p/38009

     

    I tested one of my data flows that had the same requirements you have (select all columns, but convert the timestamps) using both Redshift and MySQL and I did see both more consistent run times and a lower 30 day average duration when using MySQL.

     

    So you might want to make a copy of your data flow & convert it to MySQL to see if you can achieve the results you are looking for.

Answers

  • DDalt
    DDalt Member
    Answer ✓
    Options

    I was curious about Redshift run times being inconsistent and posed a question on DOJO day. Here is the question and response:

     

    https://dojo.domo.com/t5/Beast-Mode-ETL-Dataflow/Redshift-vs-MySQL-vs-ETL/m-p/38009

     

    I tested one of my data flows that had the same requirements you have (select all columns, but convert the timestamps) using both Redshift and MySQL and I did see both more consistent run times and a lower 30 day average duration when using MySQL.

     

    So you might want to make a copy of your data flow & convert it to MySQL to see if you can achieve the results you are looking for.

Categories