Issue raised by snowplow/snowplow-mini#378. This issue relates only to the nsq source.
To improve Mini performance we want to use larger batches when sending events to Elasticsearch. But with low event volumes, we don't want events to get stuck in the buffer, while waiting for the buffer to fill.
The solution is to have an asynchronous process that periodically flushes the buffer.
Issue raised by snowplow/snowplow-mini#378. This issue relates only to the nsq source.
To improve Mini performance we want to use larger batches when sending events to Elasticsearch. But with low event volumes, we don't want events to get stuck in the buffer, while waiting for the buffer to fill.
The solution is to have an asynchronous process that periodically flushes the buffer.