Lofesa commented on issue #1720:
URL: 
https://github.com/apache/incubator-pagespeed-ngx/issues/1720#issuecomment-734703419


   > Visible difference with fastcgi on: 
https://gtmetrix.com/compare/OhWm2FXv/PQSzLJfz .
   
   1.- Yes. With some type of page cache you avoid all time the background 
takes to generate the html, but you must take care on the cache refresh. What 
happend when you add a new post? Are the Archive pages refreshed?
   
   >  i wonder if gzip or brotli can help with this compression or maybe some 
settings in pagespeed? it seems lazyload isnt really working is that normal?
   
   2.- Your resources are now brotli compressed. What level of compression have 
you set? Level 6 is like level 9 for gzip.
   Images should not be compressed. To make it more ligth you need to play whit 
image [quality 
parameters](https://www.modpagespeed.com/doc/filter-image-optimize#image-quality)
 in pagespeed.
   
   > this seems to be the differences between the two https://uhl.site 
locations. I think my domain mapping and the way it loads might not be best way 
optimised.
   
   3.- But for now, you don´t know from where come these second request. Try to 
change to a default theme and test if still here. If it is, then the issue is 
in server config, if not, the the issue is in the frame you use (oxygen 
builder). 
   
   > Will this work?
   
   4.- Sure, but for make these config, you need to change all your server 
config. Most of these configurations are for using a proxy. Use of a proxy make 
sense in some infrastructure configs, like using a firewall:
   user -> proxy -> firewall -> server
   or behind a load balancer
   user -> proxy -> load balancer -> servers
   In a single domain, like you have, I think no advantage using these config.
   
   >  Also for the https I understand correctly it requires the local ssl ca 
from each OS loaded. Not the self signed or bought domain S
   
   5.- Yes, that is correct.
   
   6.- Using LoadFromFile with wordpress, you need to care for 2 things:
   
   A.- Pagespeed cache stores the optimized assets for the time they have in 
their cache-control header, but if you load a file from disk, this don´t make 
any http request, so don´t have any of the http header aka no cache-control 
header. When pagespeed stores a asset in their cache w/o cache.control-header 
the time the asset live (ttl) in the cache are 5min.
   To avoid this you need:
   [pagespeed 
ImplicitCacheTtlMs](https://www.modpagespeed.com/doc/system#implicit_cache_ttl) 
for any resource fetched by http request w/o cache-control header.
   [pagespeed 
LoadFromFileCacheTtlMs](https://www.modpagespeed.com/doc/system#load_from_file_cache_ttl)
 to set a cache-control for resources loaded from disk.
   
   B.- In wordpress you have request that are not in a file. Think in 
`https://uhl.site/hello-world/`, hello-world is not a file, so pagespeed uses 
Serf to fetch the resource (Serf is a headless browser) and is https, so you 
need to use:
    ```
    pagespeed FetchHttps enable;
     pagespeed SslCertDirectory directory;
     pagespeed SslCertFile file;
   ```
   If you OS is Debian/Ubuntu based, you only need the directory, bu if s 
redhat/centos based, you need both, directory and file
   https://www.modpagespeed.com/doc/https_support#configuring_ssl_certificates


----------------------------------------------------------------
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.

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


Reply via email to