We’re evaluting MM for a huge publisher website with 100.000+ articles and 500+ pages.
Our issue is that the performance with a large number of articles decreases drastically. Rendering 100.000 articles with a very simple template (basically just a prototype text output) takes around 2hrs with a decent setup (CPU/SSD).
Our goal would be to render single files most of the time (e.g. attached to a hook once the editor publishes or updated the article). This is already possible using the " --glob [ filename ] --no-clean" parameters, but even that takes around 5 minutes with the file scan we can see in verbose mode and “== Prerendering CSS” which stays for around 3 min on the command line.
Q1: Is MM a suitable tool for our approach?
Q2: Is anyone running MM with such an amount of articles / mostly single-file publishing approach?
Q3: How can the performance with this amount of data be increased, especially when generating just a single article?