* Update elasticsearch/kibana instructions
[From elastisearch notes](https://www.elastic.co/guide/en/elasticsearch/reference/current/important-settings.html#heap-size-settings) :
```
By default, Elasticsearch automatically sets the JVM heap size based on a node’s roles and total memory. We recommend the default sizing for most production environments.
```
* Update nginx conf to TLSv1.3 and IPv6
* Replace nginx proxy by native https server
Kibana now provide https web server, remove the nginx proxy part and directly use kibana
* Fix typo
* Add infos how to login to kibana
* Add interface details
* log the stack trace when handling an exception
* fix issue when batch size is not supplied, raise error on fetch_messages
* fix unused var
* initialize graph client with scopes,
* add check for status code
* updates `parse_aggregate_report_xml` to not raise an unhandled
exception on parsing errors
* adds an empty xml file to the aggregate test samples
* adds test for coverage
* Resolves#348
* append to output files if they already exist
instead of overwriting them. fixes#226
* save output with each successive watch result
possible now that it appends to the output
* implement token cache for Graph API
* fix case when username isn't defined (device code)
* add batch size when listing messages from Graph API
* fix string concat
- Support MFA for Microsoft Graph (PR #320 closes issue #319)
- Add more options for S3 export (PR #328)
- Provide a helpful error message when the log file cannot be created (closes issue #317)
* implement DeviceCode and ClientSecret auth flow for MS Graph
* update readme for MS Graph auth method additions
* add warning to msgraph config setup
- Support non-standard, text-based forensic reports sent by some mail hosts
- Set forensic report version to `None` (`null` in JSON) if the report was in a non-standard format and/or is missing a version number
- The default value of the `mailbox` `batch_size` is now `10` (use `0` for no limit)