So I understand now that you are charged by the time that your pod is live. But I don’t know if you can ‘stop’ or ‘pause’ an index when you don’t need to use it, without deleting the index?
Our use case is very demand driven and the index does not need to be online all the time. Typically, a couple of days will be needed to start the index, process the data, then ‘shut down’ the index.
Is this possible? If not, the price is quite expensive for us. If yes, how to do this? I could not see a way to do this through my dashboard (although I am only free tier user right now).
You could pause a paid index by turning it into a collection and then creating an index from the collection in the future. When you restart be careful to allocate sufficient resources (pod count or pod size) to the index to avoid errors.
Thank you. This seems rather complicated but at least it is a possibility…
But can you shed some light on the time it takes for such operations? I read here that creating an index from a collection takes 10 minutes (unless it’s a p2 pod), but
- Does this time increase wrt the number of vectors or the size of the index, as I have a lot of metdata data other than the vectors?
- How long does it take to save an index to collection?
Many thanks again!