Consuming sampled stream data
Consuming data from the sampled stream is very similar to the filtered stream integration recommendations. For understanding how to consume from a streaming endpoint, please follow the recommendations given in the consuming streaming data integration guide, and note the following differences with sampled stream.
- Although only 1% of Tweets, the volume of the sampled stream can be very high, and your client should have a high volume capacity.
- Tweets delivered through the sampled stream endpoint do not count towards the monthly total Tweet volume.
- Tweets delivered with sampled stream will not have a matching rules object.
When consuming realtime data, maximizing your connection time is a fundamental goal to promote both data reliability and data full-fidelity. When disconnects occur, it is important for your client to automatically detect the disconnection, handle the disconnection and reconnect quickly.