Peter Bourgon
Peter Bourgon
Ingesters for sure. We want to be able to support commodity log shippers.
OK, after Slack discussion with @fabxc, we learn that this enhancement is nice-to-have, to detect lost data at e.g. query or read repair time. With that in mind, we expect...
@fabxc Do you envision topic prefix after ULID on per-record basis, or is it not necessary?
My understanding is that topic would need to be first-order parameter of any query, initially just selecting specific topic, maybe eventually allowing querying multiple topics at once, but never anything...
What about this? (Not necessarily suggesting it, just to provoke discussion.) - If the ingesters are deployed in `-topic-mode none` mode, then all records are assumed to be part of...
The problem here is if we mandate use of `oklog forwarder` on the other side of the `ingester`, I'm convinced we make things very annoying for many people. I'm imagining...
For context can you describe the use-case where mode=dynamic is most useful? I'm struggling a bit because my assumption is that a single connection and forwarder would map to a...
Sure, I'm open to it, of course. edit: It would make sense both for query/store communication, as well as ingest/store replication.
Yep, to be explicit the communication paths are | Components | Protocol | Security | |:---:|:---:|:---:| | Forwarder–Ingester | plain socket | whatever you want | | Ingester–Store | HTTP/1.1...
Any transport abstraction beyond plain TCP or HTTP, especially something like ZeroMQ or mangos, would be severe overengineering at this stage of the project.