ASF GitHub Bot commented on METRON-1520:

Github user cestella commented on a diff in the pull request:

    --- Diff: metron-platform/Performance-tuning-guide.md ---
    @@ -43,6 +43,16 @@ parallelism will leave you with idle consumers since 
Kafka limits the max number
     important because Kafka has certain ordering guarantees for message 
delivery per partition that would not be possible if more than
     one consumer in a given consumer group were able to read from that 
    +## Parser Tuning Suggestions
    +If you are using stellar field transformations in your parsers, by 
default, stellar expressions
    +are not cached.  Turning on caching via setting the `cacheConfig` 
    +in your parser configuration can have performance impact if your stellar 
expressions are
    +complex (e.g. `ENRICHMENT_GET` calls or other high latency calls).  The 
tradeoff, though, is
    +that non-deterministic stellar expressions will yield cached results which 
may be wrong,
    +for the period of time in which the data exists in the cache (the max time 
in the cache is
    --- End diff --
    Thanks for the feedback.  I've accepted your rewordings.

> Add caching for stellar field transformations
> ---------------------------------------------
>                 Key: METRON-1520
>                 URL: https://issues.apache.org/jira/browse/METRON-1520
>             Project: Metron
>          Issue Type: Improvement
>            Reporter: Casey Stella
>            Priority: Major
> Given how important caching is in the enrichment topology, we should have 
> caching for stellar field transformations in the parsers as well.

This message was sent by Atlassian JIRA

Reply via email to