wildcard doesn't get appended to my ccs index pattern *:makelogs-

[LeeDr]

Kibana version: 6.2.2 SNAPSHOT

Elasticsearch version: 6.2.2 SNAPSHOT

Server OS version: Ubuntu

Browser version: Chrome

Browser OS version: Windows 10

Original install method (e.g. download page, yum, from source, etc.): .deb packages

Description of the problem including expected versus actual behavior:
We have Cross Cluster tests which create these index patterns;

data:makelogs-*
local:makelogs-*
*:makelogs-*

Knowing that the UI is going to add a * to the end of index patterns I'm not putting the last * char into the field. But I guess the pattern that starts with a * breaks the code that decides if it should append a *.

Steps to reproduce:

1. create a new index pattern and type *:makelogs-
2. expect *:makelogs-* but it doesn't add the trailing *

Errors in browser console (if relevant):

Provide logs and/or server output (if relevant):

[chrisronline]

This isn’t unexpected actually. If you type a * as the first character, the code won’t append another wildcard. I’ll look more but I’m not sure if I’d consider this a bug

[LeeDr]

Its unexpected from my user point of view. And it makes it difficult to code around in automated tests.
I currently have this in the integration-test repo so that if I pass in an indexPatternName which ends in a * then I leave that off of the characters I type in field.type(indexPatternName.slice(0,-1));.
But if I pass in an indexPatternName which does not end in * then I remove the one that the UI automatically appends.

  async setIndexPatternField(indexPatternName = 'logstash-*') {
    PageObjects.common.debug(`setIndexPatternField(${indexPatternName})`);
    const field = await this.getIndexPatternField();
    await field.clearValue();
    if (indexPatternName.endsWith('*')) {
      // if we passed in an indexPaternName that ends in *, remove it because the UI is going to add it
      field.type(indexPatternName.slice(0,-1));
    } else {
      // if we passed in an indexPatternName without *, remove the one the UI adds
      await field.session.pressKeys(indexPatternName.split(''));
      await field.session.pressKeys([Keys.ARROW_RIGHT]);
      await field.session.pressKeys([Keys.BACKSPACE]);
    }
    await PageObjects.header.getSpinnerDone();
  }

I could code-around this special case, but it seems like the UI should be putting an * to the right of any non-* chars I type (but still allow me do delete the *).

[chrisronline]

So maybe the code should consider the next character if you type in an initial * and if that is NOT a *, then it will do the same append logic? Will that work for you?

[LeeDr]

@chrisronline that sounds like it would work.

But I don't think it fixes the most general case. This probably isn't required often but what if I want logs*something-*. In that case the * isn't the first char but it seems like it would still make sense to append * to the end just like you would if it didn't have an embedded *.
I don't have a great solution to suggest, because you still have to support index patterns that don't end with *.

I've also had problems where the VM running my Kibana is slow and when I type logstash I get l*ogstash. So it seems latency can mess with the code that tries to append the *.

BTW. I wasn't sure what you would do with this issue or when, so I DID already work-around this in our Cross Cluster Search tests in the integration-test repo. It's not pretty but it works.

[chrisronline]

Pinging @skearns64 for thoughts here

[skearns64]

This sounds like it's mostly an automated test problem. If you were a live user, you'd see that the * wasn't being appended after you typed the leading *, and you'd be able to add it.

If we were going to try to address this now, I would vote for taking into account the syntax for cross-cluster search (e.g. *:something-*), but wouldn't try to handle embedded *'s in the index name (e.g. don't try to gracefully handle something like logs*something-*), and still add a trailing *. This is mostly a convenience for a common usage pattern, and advanced usage will require manual editing.

Personally, I expect we'll get more user feedback about the process of adding an index pattern over time, and we'll want to improve the UX, but I would vote to wait for a more clear set of feedback from a few sources and try to address the broader set of feedback.

[LeeDr]

I'm OK with that. Maybe we'll get some usability feedback at the Elastic{ON} session on it.

[chrisronline]

Cool. This thing is fairly wonky and it's not something we do often so it might always be a battle of what's more important

[elasticmachine]

Pinging @elastic/kibana-app-services (Team:AppServices)