Adlago

1. Loading = lazy makes loading icons delay, but does not onload them after loading the page, data-src should be added 2. Actual size of icons is 24 x 24, so this size should be specified in link 3. For responsive rendering, you should also add, for example, class=react, respectively, add the necessary in the custom css, and neutralize the original max sizes from the framework css. 4. Link should be edited in templates reactionBlurb and reactionOverview I use this: Link edited <img class="react" src='{expression="\IPS\Text\Parser::blankImage()"}' data-src='{$reaction->_icon->url}' alt='{lang="reaction_title_{$reaction->id}" escape="true"}' width="24px" height="24px" loading="lazy"> Custom css .react{width:28px;height:28px} .ipsReact_reactCount img {max-width: 28px;max-height: 28px;} @media screen and (max-width: 767px) { .react{width:20px;height:20px} }

Try - disable 3rd party plugins

The entire platform is a sufficiently complex software product. I.e. developers have a fairly high knowledge quotient. So it's strange that product protection is so bad... I think there are two lines of reasoning - one is for advertising purposes, i.e. the more programmers "struggle" with security codes, the more customers learn about this product. The other direction - lack of analytical algorithm to stop "nulled versions" - I think it is excluded, or ridiculously admissible... However, the truth is in the normality of development, and this normality is paid for. The other is from the devil...

Have you tried rolling back 7.x PHP?

They are probably looking for a balance between cash and reality, ie. any additional, external to the suite, cache creates an issues at some point. For example, I gave up nginx cache server because I got spikes on the first byte over 1.5 seconds and quite often.. Now I use only basic apache and these issues are gone. Also - look for a reasonable sidebar cache time - I use 2 hours, after much experimentation... Etc.

This mod is applied to "regular sites". In this case, look for other forms of optimization, but remove this mod.

Unfortunately, there is no answer that helped me. I ditched the nginx cache server, and now only use apache. I changed the cache sidebar to 1 hour, and for now the momentary spike lag issue is gone. There is probably more issues cache processing - but if I find something else I will write... I have a feeling that every newer version worsens cache processing... I hope I'm wrong...

@SJ77 Just a hypothesis - you are updating from 4.4 to 4.7.7 I'm not sure from which version service worker running was introduced, but I think it was after 4.5. Now in many tests - your landing page lags over 12 seconds because of this service. Your other pages load 10x faster.. It is possible that this new service does not cover your landing page, so this rather long wait for your landing page causes a high level of load on your server. Try it - shut down your landing page for half an hour and monitor server load... You might need a landing page edit because most of your users are guests, and that's a lot of high-latency requests...

@SJ77 I have some guesses, but if you want, send me your site link so I can do some tests. You can try this: open includeMeta template, at the bottom stop preloading fonts, or this: <!-- <link rel="preload" href="{url="applications/core/interface/font/fontawesome-webfont.woff2?v=4.7.0" base="none" noprotocol="true"}" as="font" crossorigin="anonymous"> -->

Try it - put this link in the address bar and enter. https://your site/admin/?app=core&module=support&controller=support&_new=1 This will probably load a normal support page

Is this Apache server not yours?

You are wrong - I tested all of this without any changes to your build - no difference in the offset of those first byte peaks. With my improvements at least in the time left between two peaks of the first byte, my site is well accelerated. PS. In fact, these first byte spikes exist on your site as well. Here now after about 10 minutes of tests - this is what I got (in the rest of the time I get your first byte in 30 - 50 milliseconds)

Too curious a suggestion... And how does css affect the first byte? I look forward to your discovery. @Charles Please explain to me: Why with my configuration - apache main server and nginx cache server, redis on refresh creates a spike of one second delay in first byte, at different interval - sometimes every 6-7 minutes, sometimes every 10-12 minutes? Stopped redis - switched to file storage data - same bad storage handling - peak after peak every 6.7 or 10 minutes. (I noticed a match with entries in the datastore directory) I've given up on that too, switched to mysql data storage - at least it only creates such a spike every 15 minutes or so... Please explain - why is there no option in ACP to select a refresh time, with each of these methods? My cache server refreshes every 30 minutes ie. what should I do (besides rebuilding the css as you advise...Ha Ha) to sync cache refresh?

Here is a line from that request 15:52:11 read(5, "<?php\n/**\n * @brief\t\tInitiates Invision Community constants, autoloader and exception handler\n * @author\t\t<a href='https://www.invisioncommunity.com'>Invision Power Services, Inc.</a>\n * @copyright\t(c) Invision Power Services, Inc.\n * @license\t\thttps://www.invisioncommunity.com/legal/standards/\n * @package\t\tInvision Community\n * @since\t\t18 Feb 2013\n */\n\nnamespace IPS;\n\n/**\n * Class to contain Invision Community autoloader and exception handler\n */\nclass IPS\n{\t\n\t/**\n\t * @brief\tClasses that have hooks on\n\t */\n\tpublic static $hooks = array();\n\n\t/**\n\t * @brief\tLoaded Hooks\n\t */\n\tpublic static $loadedHooks = array();\n\t\n\t/**\n\t * @brief\tUnique key for this suite (used in http requests to defy browser caching)\n\t */\n\tpublic static $suiteUniqueKey = NULL;\n\t\n\t/**\n\t * @brief\tDeveloper Code to be added to all namespaces\n\t */\n\tprivate static $inDevCode = '';\n\t\n\t/**\n\t * @brief\tNamespaces developer code has been imported to\n\t */\n\tprivate static $inDevCodeImportedTo = array();\n\t\n\t/**\n\t * @brief\tVendors to use PSR-0 autoloader for\n\t */\n\tpublic static $PSR0Namespaces = array();\n\t\n\t/**\n\t * @brief\tCommunity in the Cloud configuration\n\t */\n\tpublic static $cicConfig = array();\n\t\n\t/**\n\t * @brief\tIPS Applications\n\t */\n\tpublic static $ipsApps = array(\n\t\t'blog',\n\t\t'calendar',\n\t\t'cloud',\n\t\t'cms',\n\t\t'core',\n\t\t'downloads',\n\t\t'forums',\n\t\t'gallery',\n\t\t'nexus',\n\t\t'convert'\n\t\t);\n\n\t/**\n\t * Get default constants\n\t *\n\t * @return\tarray\n\t */\n\tpublic static function defaultConstants()\n\t{\n\t\t$storeMethod = 'FileSystem';\n\t\t$storeConfig = '{\"path\":\"{root}/datastore\"}';\n\t\t$cacheMethod = 'None';\n\t\t$cacheConfig = '{}';\n\t\t$redisConfig = NULL;\n\t\t$redisEnabled = FALSE;\n\t\t$outputCache = 'Database';\n\t\t$outputCacheConfig = NULL;\n\t\t$guestTimeout = 900;\n\n\t\tif ( isset( self::$cicConfig['guests']['guest_cache_timeout'] ) and self::$cicConfig['guests']['guest_cache_timeout'] and \\is_numeric( self::$cicConfig['guests']['guest_cache_timeout'] ) )\n\t\t{\n\t\t\t$guestTimeout = \\intval( self::$cicConfig['guests']['guest_cache_timeout'] );\n\t\t}\n\n\t\tif ( isset( self::$cicConfig['redis"..., 58552) = 58552 15:52:11 close(5) = 0

I'm not sure since when, probably since a recent version, when loading my site I've noticed intermittent large server response delays for the first byte. The period of these delays requests is about 6 minutes, sometimes 10-12 minutes. With such a delay, I half-hear a server response after a second and a half, and sometimes, but less often, more than 10 seconds. Typical first byte time for my site is in the 20-80 millisecond range. I checked a lot of things but didn't find a solution. I asked my host support for assistance, and they were able to record one such delaying request. What's impressive - in one minute this request sent 268 requests to your server https://www.invisioncommunity.com There is probably a problem with your license server Please check. I'm attaching the strace.txt file that my host saved. strace.zip

Apparently everyone has to save themselves, from ill-conceived things... It's a shame.

I've tested with 15 minutes, I've tested with 60 minutes, now I've left it at 17 minutes - but it's annoying when the site hangs for a second and more, and then it pours quickly and accurately... Is there an option with a cron task to sync?

I am using configuration - main server apache, cache server nginx. What is observed - (of course with the help of my host support) - When Redis recache data - the first byte of my site increases sharply over a second and a half. In the rest of the time, the first byte is within 150 - 450 milliseconds. Well, I turned off Redis - switched to file system.....data storage method. Oh wonder - first byte is now : 14- to 50 milliseconds...?!!? Tried different times with the other cache too - same crap - when sidebar cache is refreshed - first byte bounce again, over a second... There is no option to sync with nginx cache server, but tell me please how to turn off your cache sidebar blocks? Thanks

What does this have to do with fonts?

Why are the css rules below in the font.css file? .sr-only { position: absolute; width: 1px; height: 1px; padding: 0; margin: -1px; overflow: hidden; clip: rect(0, 0, 0, 0); border: 0; } .sr-only-focusable:active, .sr-only-focusable:focus { position: static; width: auto; height: auto; margin: 0; overflow: visible; clip: auto; }

Try creating a copy of your theme, then try downloading a copy theme.

In my practice I use this online tool - I always have excellent results. https://image.online-convert.com/convert-to-webp

Login ftp Open .../admin/applications/core/modules_admin/mycp and replace dashboard php with this dashboard.php

After {session key} add &app=forums then Enter

Possible dashboard php issue Try this - when you get this error after logging in ACP, in the browser command line, replace the command final from &app=core on &app=forums and Enter key If this action gets you into ACP, I will send you a dashboard php to replace yours with.