Category: All posts
Mar 10, 2021
Posted by
Team Timescale
A selection of FAQs from Timescale Office Hours. During our last session we discussed schema creation, hardware requirements, migrating to Timescale Cloud and more.
We had a great office hours session last week with attendees in various spaces including fintech, AI, and IoT. This month, we are hosting two more office hours on August 15 at 4pm ET and August 30 at 9am ET. You can sign up for one of these sessions here.
Below, we will share a selection of questions and answers from last week.
There are only two differences when creating a schema in TimescaleDB. The first is that any unique key you have on your hypertable needs to include a time field. The second is that you can not have a foreign key within a hypertable, but you can have one outside of it.
Yes, but you will want to set it in terms of the amount of time a chunk contains. We say your chunk sizes should fit into ¼ of your memory because with heavy operational workloads, you want to ensure that your b-trees that are being updated can fit in memory.
It depends on the type of query you are working with. In most cases, if you have a unique key or index, it will speed up the query.
Your hardware requirements will depend heavily on the types of queries and the percentage of your dataset that can fit in memory. Some of our benchmarking blog posts might be relevant to your use case (here and here), but otherwise, we suggest getting in touch with us via Slack to discuss further.
Yes! You can use pg_dump and restore your data on Timescale Cloud using a tool such as psql. In the future, within Timescale Cloud you can upgrade your instance with just one click and minimal downtime.
Stay tuned for more posts like this one and check out our previous recap. We hope to see you at an upcoming office hours!