Great walkthrough. I would be really interested to know what the most cost-effective alternative is to Community Cloud, especially when making it available to a large audience.
Thanks, I'm glad it was helpful. I'm going to follow this up by sharing a real-world example, which will be a better context for discussing the strengths and weaknesses of Community Cloud. For example there's some optimizing you can do around caches and UI decisions that affect how resource-intensive the deployed app is. But you don't have any metrics around usage other than number of visitors, so the only feedback is pretty much it works or it doesn't. You can't tell when you're running up against resource limits.
I think I'm going to deploy to a VPS just to see more fine-grained resource usage stats, and measure the impact of a variety of decisions. One recommendation Streamlit makes is only having "dozens at most" of input variations; ie the matrix of all possible input settings. That makes sense for some use cases, but it's really limiting for others.
I think this really helps clarify what projects are suitable for a framework that does all the calculations in Python, vs one that offloads the calculation work to the client via JavaScript.
Great walkthrough. I would be really interested to know what the most cost-effective alternative is to Community Cloud, especially when making it available to a large audience.
Thanks, I'm glad it was helpful. I'm going to follow this up by sharing a real-world example, which will be a better context for discussing the strengths and weaknesses of Community Cloud. For example there's some optimizing you can do around caches and UI decisions that affect how resource-intensive the deployed app is. But you don't have any metrics around usage other than number of visitors, so the only feedback is pretty much it works or it doesn't. You can't tell when you're running up against resource limits.
I think I'm going to deploy to a VPS just to see more fine-grained resource usage stats, and measure the impact of a variety of decisions. One recommendation Streamlit makes is only having "dozens at most" of input variations; ie the matrix of all possible input settings. That makes sense for some use cases, but it's really limiting for others.
I think this really helps clarify what projects are suitable for a framework that does all the calculations in Python, vs one that offloads the calculation work to the client via JavaScript.
Thanks, I look forward to the follow up article.