jmarantz commented on issue #2089: URL: https://github.com/apache/incubator-pagespeed-mod/issues/2089#issuecomment-1095540226
Thanks for the data including the config; it's very helpful. It does look like the performance is pretty good even from the US, despite the TTFB. I assume the server is in Germany with no US CDN. Of course TTFB will never get faster with PageSpeed, only slower. The question is whether the optimizations it makes during that time provide a better user experience. I visited view-source:https://www.keinegotterkeinemeister.de/tshirt-sabotage-black-cat-D01004174288P0210/?PageSpeedFilters=+debug to see what timing info PageSpeed injected at the end, and found this: ``` #NumFlushes 0 #EndDocument after 5906us #Total Parse duration 5685us #Total Render duration 411563us #Total Idle duration 221us No critical images detected. The following filters were disabled for this request: CriticalSelectorFilter: No critical selector info in cache DelayImages ``` So it looks like your server spent 5.7 seconds in PageSpeed parsing your HTML -- that's mostly CPU bound. Then it spent 452ms rendering it -- that may be a combination of image optimization, network operations, waiting for asynchronous optimizations and URL fetches (depending on the cache setup) That leaves 1100-270= 378ms unaccounted for. I'm guessing that most of this comes from the file cache. It's possible your disk is I/O bound. My main suggestion is to try deploying memcached or redis as an asynchronous cache. That should be faster than the file-cache. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: dev-unsubscr...@pagespeed.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org
