bug/Set a longer timeout for ES integration tests using allegro

[ghost]

Among the many issues spotted when running Elasticsearch integration tests on the CI environment, there is one that has not been addressed yet, and is failing some of the tests with errors.

See empty pull requests #2237 and #2238 (used to test whether concurrency of ES tests works now that the installation directory is specified).

The issue at hand is a timeout to actually spin up the Elasticsearch runtime at Allegro's level.

This experimental task is about trying to specify a much larger timeout to accommodate with the machine's limitations.

[ghost]

Looks like the jenkins builds on the two concurrent "empty" PRs I used for testing this is green. Waiting for finalization of the travis build in case something else comes up - otherwise we might have a winner here finally (although only continued usage will tell if those measures were good enough).

[ghost]

Merged into concurrent empty PRs and tested building concurrently again - all green for now, so closing with hope.
We can still iterate later if 1' not enough.
See also #2229.