Refined retention/replication policy: towards a decentralized timeseries platform

Hi all!

This is might be an enhancement request for the retention policy.

We have a cluster of InfluxDB nodes, where one particular node is ingesting series at a throughput higher than these series can replicate/spread over the cluster. We want to specify that some of theses series shall be retained locally, analyzed, then dropped after some duration. Sometime, we want to change the retention policy of some interesting analyzed series to make them available to the other nodes. Furthermore, the ingesting node is frequently disconnected from the other nodes of the cluster for several days, so the application of retention policies shall be robust to network cut.

How could we satisfy our usecase with the current or future version of TICK stack ?

Best regards,

@Lionel What do you mean by a cluster of Influx nodes? Are you talking about InfluxData Enterprise?

@jackzampolin Right now we are only using the free InfluxDB version, for evaluation purpose, but yes I mean “InfluxData Enterprise” if it is the only way to create a cluster. Do you want me to modify my post?