I am looking into ways to monitor and report our instance’s usage over time.
I noticed this post (Monitoring CPU, Memory and GPU Usage) with a great suggestion for hardware monitoring, and plan to use that to take care of the node itself.
However, I am interested in what my users are doing in the instance itself - in projects, workspaces, and jobs. Is there any way to export summaries or metadata of things like this to look at usage trends?
Thanks for the post! This is a great question. In an upcoming version of cryoSPARC we’ll have better support for viewing aggregate statistics in a user friendly manner. For now I can provide a Python script you can run in the terminal that can pull some of these statistics if you’re interested.
Because project, workspace and job metadata is stored in MongoDB, it makes it very easy to run ‘aggregation pipelines’ to summarize information about the cryoSPARC instance.
Could you give me a few examples of exactly what type of information you are interested in? For example, type and number of jobs completed by day/week/month/quarter, average runtime of jobs by type, instance usage by user, etc.