Fix directly by increasing client default timeout to a higher (configurable) value.
Also, optimize twitter_monitor by reducing chattiness with Elasticsearch - instead of indexing each new tweet in a separate request, collect them in a bucket and make a bulk index request when the bucket is full.
Fix directly by increasing client default timeout to a higher (configurable) value.
Also, optimize twitter_monitor by reducing chattiness with Elasticsearch - instead of indexing each new tweet in a separate request, collect them in a bucket and make a bulk index request when the bucket is full.