Disable default geoip database downloading for Elasticsearch container

[lhotari]

Motivation

By default, Elasticsearch container will download the geoip lite database.
It's better to disable this behavior by default.
There's also an elastic issue about this default behavior: elastic/elasticsearch#90673. Fixed since v8.7.0 with elastic/elasticsearch#92335 .

Modifications

geoip downloading can be disabled by passing -Dingest.geoip.downloader.enabled.default=false in JVM options or passing ingest.geoip.downloader.enabled=false in the environment.

The problem with passing ingest.geoip.downloader.enabled=false in the environment is that it's not compatible with OpenSearch. OpenSearch will fail to start with a configuration validation error org.opensearch.bootstrap.StartupException: java.lang.IllegalArgumentException: unknown setting [ingest.geoip.downloader.enabled] please check that any required plugins are installed, or check the breaking changes documentation for removed settings. Some users might be using ElasticsearchContainer class together with the OpenSearch docker image.

[eddumelendez]

JFTR, OpenSearch provides its own module

[lhotari]

The elastic issue description elastic/elasticsearch#90673 is interesting:

In Elasticsearch 7.14 a feature was added which would automatically download the MaxMind GeoIP database. Specifically, it downloads the database from Google Cloud after fetching this JSON object. This is enabled by default.

We discovered that this feature was responsible for around $20,000 on our AWS bill since we upgraded last March and based on a rough approximation of traffic we were charged for it probably cost Elastic about $5,500 💸 Investigation was slow and required enabling AWS VPC logs to assess the NAT gateway traffic source.

[lhotari]

I guess this geoip download doesn't impact the default image tag that ElasticsearchContainer sets, 7.9.2. The feature was introduced in 7.14.0.
7.14.0 release notes includes:

• Enable GeoIP downloader by default #74323

[eddumelendez]

Thanks again, @lhotari ! The context is really useful, let's keep compatibility in case OpenSearch is used along with ElasticsearchContainer