7.2.0 release notes

[jakelandis]

highlights are still a todo..

Due to the timing and state of branches at time of the label creation the initial attempt to create these resulted in many false positives. To get a more correct list I added all of the released 7.x versions to be ignored in the script that generates these.

my %Ignore = map { $_ => 1 }
    ( ">non-issue", ">refactoring", ">docs", ">test", ">test-failure", ">test-mute", ":Core/Infra/Build", "backport",  "v7.0.0", "v7.0.0-beta1", "v7.0.0-rc1", "v7.0.0-rc2", "v7.1.0", "v7.1.1",  "v7.0.1" );

After I had the more refined list, I cleaned up the [category] in the commit message and other other minor spelling and grammar issues.

Finally, I manually moved some items between categories. For example, I moved some new features down to enhancements and vice-versa. This step is mostly subjective and would appreciate a second opinion that the categories are are correct. (I didn't move any of the bugs around).

Highlights will be added on this PR or different PR's as I coordinate them independently.

[elasticmachine]

Pinging @elastic/es-docs

[jakelandis]

@elasticmachine run elasticsearch-ci/docbldesx

[jakelandis]

Hmm... issues talking with Github:

13:09:49 Cloning into 'docs'...
13:11:57 fatal: unable to access 'https://github.com/elastic/docs/': Failed connect to github.com:443; Connection timed out

trying again.
@elasticmachine run elasticsearch-ci/docbldesx

[lcawl]

Hmm... issues talking with Github:

These issues are not limited to your PR. I was able to build the Elasticsearch Reference and the Installation and Upgrade Guide (which re-uses the breaking changes) successfully.

[lcawl]

@droberts195 I incorporated the ml-cpp PRs from https://github.com/elastic/ml-cpp/blob/7.2/docs/CHANGELOG.asciidoc here

[lcawl]

LGTM and builds successfully

[droberts195]

This change had both a Java and a C++ component. The Java change is release noted as "Increase maximum forecast interval to 10 years" 8 lines above.

What do you think about combining them @lcawl. Maybe delete the two current items and replace with something like:

• Increase maximum forecast interval from 8 weeks to a limit based on the amount of data seen, up to a maximum of 10 years {ml-pull}214[multicast fails with network interfaces instead of IP #214] and {pull}41082[Increase maximum forecast interval to 10 years. #41082] (issue: {issue}41103[[ML] Extend ML Forecast beyond 8 weeks in the REST API #41103])

But I'll leave the exact wording up to you.

[lcawl]

Thanks @droberts195 I committed that change.

[colings86]

LGTM (after addressing others comments)

[giladgal]

LGTM

[jakelandis]

@giladgal @lcawl - Could I request an additional review for the release highlights ? https://github.com/elastic/elasticsearch/pull/43343/files#diff-3e5d6e4c920b36c3bf921860e6c2b57d ?

[lcawl]

This sentence is quite long and a bit unclear:

As soon as an index is closed, Elasticsearch takes care of safely tearing down the "opened" shards before reinitializing them as "closed" shards, which require much less resources while still providing the ability for shards to be promoted as primaries or to be recovered from peer.

Perhaps it could be changed to something like this?:

As soon as an index is closed,
Elasticsearch takes care of safely tearing down the "opened" shards before
reinitializing them as "closed" shards, which require fewer resources. Closed shards can later be promoted to primary shards (link to how/when this happens?) or automatically recovered during {ref}/recovery.html[peer recovery].

[jrodewig]

Added a few nits but nothing blocking aside from @lcawl's changes.

[jakelandis]

@tlrx @ywelsch - any suggestions for Lisa's comment ? #43343 (comment) ?

[giladgal]

Other things I would consider including:
Replicated closed indices
Snapshot repositories app
SQL Median Absolute Deviation
SQL CASE/WHEN/END
HTML Strip Processor

[jakelandis]

Replicated closed indices (already covered)
Snapshot repository app - since this is a Kibana and these are ES release notes probably best for blog

SQL * and Html Strip Processor - possibly... these are great features, but trying to keep the list relatively small. We have the full release notes which covers these, and IMO the blogs can go into quite a bit more detail.

[tlrx]

@tlrx @ywelsch - any suggestions for Lisa's comment ? #43343 (comment) ?

Lisa made a good suggestion, we should change this to what was suggested. Do you prefer me to dot it or can you do it yourself as part of this PR?

Requested change made

[jakelandis]

thanks @tlrx and @lcawl - I have made the changes requested. However, I couldn't find relevant documentation to link.

[jakelandis]

@elasticmachine run elasticsearch-ci/docs-check