Joshua Marantz

I think it should be straightforward to write MultiAsyncFetch (maybe call it RacingAsyncFetch?). It should be easy to create that in isolation and unit-test it vigorously. It would buffer separately...

As mentioned above, there is a timeout you can adjust to make this happen less: BlockingFetch has an undocumented timeout (default 5 seconds, settable in pagespeed.conf). There is a TODO...

I believe this issue is fixed in 1.14. Can you upgrade?

Based on the numbers we've seen, Brotli is now in good enough shape we can deploy it. However we are pretty tight on human resources now and it is feasible...

I don't think you should need to flush the pagespeed cache when changing options. A hash of all the options are incorporated into the cache keys. However I think it...

I think that makes sense. PageSpeed doesn't know anything about Brotli right now so if it is configured to cache gzipped content and send it to clients, the Brotli module...

That is an interesting idea, and I can see the potential value. But the system isn't quite designed for that use-case. The reason, I'm guessing, that you'd prefer this architecture...

One clarification: even if you use a shared cache like memcached or redis, you need to purge the cache on each server. This is because purging the cache doesn't actually...

1. I think that's right: 1MB. That was based on limitations in older memcached and you can configure that higher or off I think. No limit on redis. 2. Yes...

I lost track of the recommendation about caches and where that came from. If you have any image or CSS optimizations, or URL-rewrirting enabled, you must have caches in mod_pagespeed...