Add WARN-level logging for `data_services_api` retry attempts

[joescottdave]

Background

This app does not currently log anything when the
DataServicesApi::Service Faraday retry middleware retries a request: if a retry eventually succeeds, that
attempt is invisible in the logs.

Retries are now configurable via retry_block options after adopting the latest
data_services_api version that added HTTP retry support (ppd-explorer: #343).

Task

• In the app's DataService model, pass connection_failed_retry_options and
  timeout_retry_options to DataServicesApi::Service.new, each with a retry_block
  that logs a WARN-level line per retry attempt (app name, exception class/message,
  attempt number, delay before retry, request path).
• Follow the org JSON logging conventions (see logging standard / Ops wiki) for field
  names: use request_status: 'processing' and path to match the existing
  ERROR-level log shape from ApiPrometheusSubscriber.
• Add a unit test on the retry_block (see ukhpi's test/models/concerns/data_service_test.rb
  for reference).
• Update CHANGELOG.md.
• Capture example log output for a real retry (e.g. via rails runner reproduction)
  to confirm the shape in Kibana.

Reference

• ppd-explorer implementation target: app/models/data_service.rb
• ppd-explorer's prior ERROR-level logging work: chore: add ERROR-level logging for api connection failures in ApiPrometheusSubscriber #339
• ppd-explorer's dependency bump enabling retries: deps: adopt the latest version of data_services_api to pick up ability to configure HTTP retries #343
• ukhpi's equivalent implementation (reference): app/models/concerns/data_service.rb, test/models/concerns/data_service_test.rb, #615, #623
• epimorphics/ops#596