Right now there is a single instance of DB node. There is an issue of this single instance being unreliable, therefore losing data.
One question is how long we continue to use single instance PromScale/TimescaleDB.
0.1 version will be using this
Then prefer PG instance only server as a cluster-local buffering storage, i.e., the data is supposed to be fetch or pushed to cloud storage of SaaS, or user's own cloud storage backend.
Will we move to multi-node PG for reliability
Might be, but it could be solved by real-time streaming by API server from Postgres storage to Cloud storage.
Right now there is a single instance of DB node. There is an issue of this single instance being unreliable, therefore losing data.
One question is how long we continue to use single instance PromScale/TimescaleDB. 0.1 version will be using this Then prefer PG instance only server as a cluster-local buffering storage, i.e., the data is supposed to be fetch or pushed to cloud storage of SaaS, or user's own cloud storage backend.
Will we move to multi-node PG for reliability Might be, but it could be solved by real-time streaming by API server from Postgres storage to Cloud storage.