The graphite-raintank finder uses two separate queries to cassandra. 1 for finding leaf_nodes and one for finding branches. When making calls to /render, only leaf nodes are used. To significantly improve rendering performance and reduce load on our Elasticsearch cluster we should skip searching for branches when rendering datapoints.
The graphite-raintank finder uses two separate queries to cassandra. 1 for finding leaf_nodes and one for finding branches. When making calls to /render, only leaf nodes are used. To significantly improve rendering performance and reduce load on our Elasticsearch cluster we should skip searching for branches when rendering datapoints.