Log Message:
Update databases/elasticsearch to 5.3.0

=== Breaking changes

Query DSL::
- Fix name of `enabled_position_increments`

Search::
- Remove DFS_QUERY_AND_FETCH as a search type

=== Breaking Java changes

CRUD::
- Make index and delete operation execute as single bulk item
  (backport of #21964)

Internal::
- Replace SearchExtRegistry with namedObject
- Replace Suggesters with namedObject

Network::
- Remove HttpServer and HttpServerAdapter in favor of a simple
  dispatch method

Plugins::
- Deguice rest handlers

Search::
- Remove QUERY_AND_FETCH search type

=== Deprecations

Core::
- Deprecate lenient booleans

Geo::
- Geo distance range deprecation

Java API::
- Add BulkProcessor methods with XContentType parameter

Mapping::
- Deprecate the include_in_all mapping parameter

Network::
- Deprecate Netty 3
- Deprecate Netty 3

Plugin Repository Azure::
- Deprecate auto creation of container for azure repository

Plugin Repository S3::
- S3 Repository: Deprecate specifying region
- S3 Repository: Deprecate auto creation of s3 bucket for repository
- Deprecate specifying credentials through env vars, sys props, and
  remove profile files

REST::
- Deprecate ldjson support and document ndjson for bulk/msearch
- Optionally require a valid content type for all rest requests with
  content

Scripting::
- Change Namespace for Stored Script to Only Use Id

=== New features

Aggregations::
- Initial version of an adjacency matrix using the Filters aggregation

CRUD::
- Allow an index to be partitioned with custom routing

Highlighting::
- Integrate UnifiedHighlighter

Search::
- Add federated cross cluster search capabilities
- Add field collapsing for search request

Settings::
- Add infrastructure for elasticsearch keystore

=== Enhancements

Allocation::
- Include stale replica shard info when explaining an unassigned
  primary

Analysis::
- Analyze API Position Length Support

CRUD::
- Make index and delete operation execute as a single bulk item

Circuit Breakers::
- Add used memory amount to CircuitBreakingException message (#22521)

Cluster::
- Connect to new nodes concurrently

Core::
- Simplify write failure handling (backport of #19105)
- Simplify ElasticsearchException rendering as a XContent
- Remove setLocalNode from ClusterService and TransportService

Engine::
- Replace EngineClosedException with AlreadyClosedExcpetion

Index APIs::
- Indexing: Add shard id to indexing operation listener
- Better error when can't auto create index

Ingest::
- Lazy load the geoip databases

Internal::
- Improve connection closing in `RemoteClusterConnection`
- Remove some more usages of ParseFieldMatcher
- Introduce ToXContentObject interface

Java API::
- prevent NPE when trying to uncompress a null BytesReference

Java High Level REST Client::
- Add parsing from xContent to MainResponse
- Parse elasticsearch exception's root causes
- Add parsing method to BytesRestResponse's error
- Add parsing method for
  ElasticsearchException.generateThrowableXContent()
- Add fromxcontent methods to delete response
- Add parsing methods for UpdateResponse
- Add parsing from xContent to InternalSearchHit and
  InternalSearchHits
- Add fromxcontent methods to index response

Java REST Client::
- move ignore parameter support from yaml test client to low level
  rest client
- Support Preemptive Authentication with RestClient

Logging::
- Expose logs base path

Mapping::
- Improve error message for ipv6 on legacy ip fields

Network::
- Keep the pipeline handler queue small initially

Packaging::
- Introduce Java version check

Plugin Discovery EC2::
- Read ec2 discovery address from aws instance tags

Plugin Lang Painless::
- Generate reference links for painless API
- Painless: Add augmentation to String for base 64

Plugin Repository S3::
- S3 repository: Add named configurations

Plugins::
- Add the ability to define search response listeners in search plugin

Query DSL::
- QueryString and SimpleQueryString Graph Support
- Additional Graph Support in Match Query
- RangeQuery WITHIN case now normalises query

Reindex API::
- Increase visibility of doExecute so it can be used directly
- Improve error message when reindex-from-remote gets bad json

Scripting::
- Expose multi-valued dates to scripts and document painless's date
  functions

Search::
- Add a setting to disable remote cluster connections on a node
- First step towards separating individual search phases
- Add parsing from xContent to SearchProfileShardResults and nested
  classes
- ProfileResult and CollectorResult should print machine readable
  timing information

Settings::
- Improve setting deprecation message
- Add secure settings validation on startup
- Allow comma delimited array settings to have a space after each
  entry
- Allow affix settings to be dynamic / updatable
- Allow affix settings to delegate to actual settings
- Make s3 repository sensitive settings use secure settings

Snapshot/Restore::
- Duplicate snapshot name throws InvalidSnapshotNameException
- Use general cluster state batching mechanism for snapshot state
  updates

Stats::
- Add geo_point to FieldStats

=== Bug fixes

CRUD::
- Fix backport executing ops as single item bulk

Cache::
- Invalidate cached query results if query timed out

Cluster::
- Don't set local node on cluster state used for node join validation
- Allow a cluster state applier to create an observer and wait for a
  better state
- Cluster allocation explain to never return empty response body

Exceptions::
- Stop returning "es." internal exception headers as http response
  headers

Ingest::
- Improve missing ingest processor error
- fix date-processor to a new default year for every new pipeline
  execution

Internal::
- Always restore the ThreadContext for operations delayed due to a
  block
- Fix handling of document failure exception in InternalEngine
  (backport #22718)

Java REST Client::
- RestClient asynchronous execution should not throw exceptions

Mapping::
- Switch include_in_all in multifield to warning
- Fix MapperService StackOverflowError
- Range types causing `GetFieldMappingsIndexRequest` to fail due to
  `NullPointerException` in `RangeFieldMapper.doXContentBody` when
`include_defaults=true` is on the query string
- Disallow introducing illegal object mappings (double '..')

Network::
- Respect promises on pipelined responses
- Ensure that releasing listener is called

Packaging::
- Fall back to non-atomic move when removing plugins

Parent/Child::
- Add null check in case of orphan child document

Plugin Ingest Attachment::
- Remove support for Visio and potm files
- Add missing mime4j library

Plugin Lang Painless::
- Fix Bad Casts In Painless
- Don't allow casting from void to def in painless
- Fix def invoked qualified method refs

Plugins::
- Add check for null pluginName in remove command

Query DSL::
- Fix parsing for `max_determinized_states`

REST::
- [API] change wait_for_completion default according to docs
- HTTP transport stashes the ThreadContext instead of the
  RestController
- Ensure we try to autodetect content type for handlers that support
  plain text
- Fix date format in warning headers
- Correct warning header to be compliant
- Fix search scroll request with a plain text body
- Handle bad HTTP requests
- Properly encode location header

Reindex API::
- Reindex: do not log when can't clear old scroll
- Fix reindex-from-remote from <2.0

Scripting::
- Remove unnecessary Groovy deprecation logging
- Script: Fix value of `ctx._now` to be current epoch time in
  milliseconds

Search::
- Honor max concurrent searches in multi-search
- Avoid stack overflow in multi-search
- Replace blocking calls in ExpandCollapseSearchResponseListener by
  asynchronous requests
- Ensure fixed serialization order of InnerHitBuilder

Settings::
- Fix merge scheduler config settings
- Settings: Fix keystore cli prompting for yes/no to handle console
  returning null
- Expose `search.highlight.term_vector_multi_value` as a node level
  setting
- NPE when no setting name passed to elasticsearch-keystore

Similarities::
- Fix similarity upgrade when "default" similarity is overridden

Stats::
- Avoid overflow when computing total FS stats
- Handle long overflow when adding paths' totals
- Fix control group pattern
- Fix total disk bytes returning negative value

Task Manager::
- Fix hanging cancelling task with no children
- Fix broken TaskInfo.toString()

=== Regressions

Core::
- Source filtering: only accept array items if the previous include
  pattern matches

=== Upgrades

Internal::
- Upgrade to Lucene 6.4.1.