We see an issue sometimes where when an image (e.g., for one of the time series charts) updates server-side due to new data, the client still loads an old version of the image from the browser cache. For the most part, we do want to continue benefiting from browser caching of data, so here's a potential idea:
Calculate a md5 or such hash of the content of the file.
Add a url parameter to the content that is based on that hash, so that it automatically changes when the content changes. This will force a reload of that item if and only if the content is expected to have changed, so in the cases where the underlying content hasn't changed, we'll still end up using the old version of the item without making a network call.
We see an issue sometimes where when an image (e.g., for one of the time series charts) updates server-side due to new data, the client still loads an old version of the image from the browser cache. For the most part, we do want to continue benefiting from browser caching of data, so here's a potential idea: