First question: Do you really need a cluster? We were running InfluxDB with a lot more data and ran on a single node. A single machine is surprisingly scalable.
If you are positive that you need a cluster, here are my suggestions: * Recently Elasticsearch has become a lot better at time series data, so it might be worth a try. * My second bet would be Cassandra. Look at these slides for more info: http://de.slideshare.net/patrickmcfadin/time-series-with-apa... Not sure if this is supported by Grafana, though. * Another one would be Prometheus which also works with Grafana. Personally I'm not a big fan of its query language. Also clustering support is pretty limited there (and mostly needs to be handled by the client).