Could you kindly share the benchmark figures for CPU and memory usage in timeseries collection? Additionally, I would like to know if any tests have been conducted to compare CPU and memory utilization between timeseries collections and regular collections.
Thank you in advance.
Welcome back to the MongoDB Community forums
We strive to minimize CPU and memory usage, but it will be difficult to compare actual benchmark numbers as they may vary depending on the specific use case.
In terms of time series vs. regular collection, these are two distinct features with different characteristics. For example, time series does not have a unique index, and therefore, comparisons can only be made for a specific workload.
Additionally, there is a difference in disk usage between the two types of collections. I suggest referring to the article “Analyzing Data Storage: Regular Collection vs Time Series Collection” for further information on this topic.
May I ask if you have experienced any performance issues with time series collection in your workload? Have you conducted any tests that demonstrate this?
Thank you for getting back to us. At this point, we haven’t run any tests, but we’re seeking to determine whether migrating from a regular collection to a timeseries collection would result in increased CPU and/or memory usage?
I would recommend testing it as it will vary depending on factors such as the volume of data, the complexity of queries and aggregations getting performed on the data, and the specific implementation of the time series collection.
However to read further please refer to the Time Series Compression and regular collections compression documentation.