Introduce a sleep before reconnecting in case connection got lost, to avoid spammy error logs

[detobel36]

Currently it is ElasticSearch library that try again to connect to database (due to parameter retry_on_timeout set on True).
So to add timer between each connection test, two solution are possible:

• wait that ElasticSearch library implement this function
• remove this auto reconnect and manage it manually

There is maybe a possibility to fetch log and only display a small message (and not all the trace)

EDIT: Linked to this: https://github.com/NVISO-BE/ee-outliers/blob/master/app/helpers/es.py#L52 EDIT2: Interesting stuff about that: https://github.com/elastic/elasticsearch-py/issues/666

[daanraman]

Elasticsearch reconnections are now wrapped in a 1 minute sleep cycle to avoid this issue.

Example:

    # Initialize Elasticsearch connection
    while not es.init_connection():
        time.sleep(60)

[daanraman]

Tested this in a local deployment where the connection fails - sleep is respected.