The only way so far is the one I described there: through developer tools. That’s where you find the sizes of 0.json and 1.json uncompressed. There has been talk of making it easier to analyze through coda UI but no commitments yet.
There’s no exact definition of “big”. There are two hard thresholds: 125 MB after which Coda daemons reject processing the doc, and 5 minutes of loading time after which Coda will refresh the tab and start loading your doc from the start if it hasn’t finished loading it already (thus possibly running you into endless loop of loading). Other than that, the “big” is only as big as your patience to wait for it to load/recalculate… and your RAM too.
This may help you. That’s how you can get image links to then bulk download: