Skip to content

Conversation

@ruflin
Copy link
Contributor

@ruflin ruflin commented Oct 19, 2018

All the original fields are not indexed. For consistency the same should be true for user_agent.original.

All the original fields are not indexed. For consistency the same should be true for user_agent.original.
@ruflin ruflin added the review label Oct 19, 2018
Copy link
Contributor

@webmat webmat left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Why turn indexing off here?

Sure it may not be the most often used field, but I can still see use cases where kw indexing can be useful.

  • exact match filtering on the whole user agent string
  • aggregations on typically shorter UA strings from various programming languages, which are often not well supported by UA parsers

* Remove `log.offset` and `log.line` as too specific for ECS. #131
* Remove top level objects `kubernetes` and `tls`. #132
* Remove `*.timezone.offset.sec` fields as too specific for ECS at the moment. #134
* Set indexing to false for `user_agent.original`.
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Why turn indexing off here?

Sure it may not be the most often used field, but I can still see use cases where kw indexing can be useful.

  • exact match filtering on the whole user agent string
  • aggregations on typically shorter UA strings from various programming languages, which are often not well parsed by UA parsers

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think if we index it by default, it should be text instead of keyword. Not indexing it by default I would hope it saves on space as the user_agent can be pretty extensive.

We can do the same trick here as we do for the keywords. Not index it by default and when someone needs it, he can add .text as a multifield.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think turning off indexing here is premature optimization and will cause problems down the line.

At the very least, though, I would change the disabled type to keyword, not text. We need to stop the text indexing on the canonical field.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

For reference, also had a discussion around this here: elastic/beats#8537

For the text type: My assumption so far was that the type does not matter if we don't index. Would be good to know if it matters.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@kvch For awareness

@ruflin
Copy link
Contributor Author

ruflin commented Dec 7, 2018

Closing in favor of #240

@ruflin ruflin closed this Dec 7, 2018
@ruflin ruflin deleted the original-non-index branch December 7, 2018 13:39
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants