Though there are far and wide systems for live debugging, and the sort, for forensic or load analysis, our php-extjs framework did not have anything pre planned, other than some query loggers, and background processing systems to log into db etc. While recently the Master MySQL server started showing variations in the cacti patterns (normal was about 4 to 20 in working window, but was steady between 35 and 40 in the tantrum period), we started to worry and could not identify the situation. Also restarting all application servers and clearing the session store would immediately drop the MySQL fsync graph to a standard pattern. This is the time when I looked for a input logger for Apache, and found about the dumpio, but needed the webserver to be restarted. Actually the time was ripe that the application was in a tantrum, and the MySQL graphs showing about 35 fsyncs average.
Revisiting Importance of event logging in server side scripting and other articles on the net, the out come was a class with a single static method. This was designed to pick and log any input. This was later moved as the lite version, and a full version capable of capturing the output also was built.
I was digging through my old codes to get this, still could not find. Hence when I wrote it, just made a note here such that later it wont be difficult to find atleast for me.
Recently in a discussion on linkedin PHP webservice – Logging Requests/Responses, Roy de Kleijn had asked “how can I print the XML in XML structure, in stead of a single line of text”, which triggered me a thought. This may be trivial for veterans, and still a new information for newbies.
Recently I got fed up trying to convince the operators of a client, for whom we had implemented a wordpress site. The operators, even after we repeatedly asked to upload only relevant sizes for certian positions, were trying to make one size fit for all. Hell, places where thumbnails of 1 to 2K would suffice, were being used up by 30K images, resized by our theme constraints. We were running wordpress on lighttpd.
After about an year, since we launched Asianet News Portal the home page with all the complications had started to show a dead slow performance, and by using our Query Profiler, I could identify the bottleneck as the number of queries, and wordpress itself was telling the status as 48 queries in 6.344 seconds. Now this was a whopping value, and when the latency of 2 to 3 seconds is added to this, the initial html load would be in 9 seconds. Making the total site complete load in about 53 seconds. One can imagine the agony I was feeling, after all the work I had done to get backlinks, and if users who are visiting the site is just running away without waiting for the site to load, then I better not run the site atall.
Recently while working for a wordpress plugin, though there are a lot of solutions for this, I could not find the exact thing which I needed, and thought about writing one of my own. Just writing it and using in the current project would be a waste, so thought I would slap it on here. It would help me in the long run.. Read more…
Recently while developing systems which consumes webservices, after a lot of optimizations were put into our code, we resorted to do some sort of benchmarking. From the begenning itself, we had proper loggers which were logging the activities to database tables, and verbose text files. For benchmarking of the systems we use xdebug, and our Open PHP Myprofiler. The code coverage analysis by xdebug, as well as cache grind showed our bottleneck was our webservice provider. While checking out, we also found that we could reduce the waiting time for webservices by requesting smaller data set from the webservice.
We were worried, about all the comments on kerala online, being marked as spam by the akismet plugin. When on detailed examination, we found that the basic problem was that wordpress was logging only the immediate downsteam ip as the remote address, well ours was a bit confusing setup, but to handle the traffic we needed it that way.
A search for wordpress behind reverse proxy, landed me to the wordpress support page. In fact the 5th entry on that page is done by Gopka, who is the lead on this project from Saturn.
We started to correct the remote address by overriding the global variable making slight changes to the wp-config, such that we will not accidentally overwrite the changes while upgrading wordpress. Well the code
$list = explode(',',$_SERVER['HTTP_X_FORWARDED_FOR']);
$_SERVER['REMOTE_ADDR'] = $list;
when added to the wp-config, it started to log the first IP, and that would be mostly private IPs if the request was from organisations where internet was shared through proxies or using NAT. The case was same at our office, so we had to find the first public ip from the list of IPs and the code for function getIpBehindProxy was the out come.
A PHP accelerator is an extension designed to boost the performance of software applications written using the PHP programming language. Most PHP accelerators work by caching the compiled bytecode of PHP scripts to avoid the overhead of parsing and compiling source code on each request (some or all of which may never even be executed). For best performance, caching is to shared memory with direct execution from the shared memory and the minimum of memory copying at runtime. A PHP accelerator typically reduces server load and increases the speed of PHP code anywhere from 2-10 times, depending on factors such as the inherent execution time of the PHP application and the percentage of source code actually executed on a given request. While a code optimizer may even slow down overall performance when used in isolation, it can provide an additional performance boost when coupled with a code cache as the optimization effort is performed just once.
Now the question was about a suggestion to have a balance between best designs considering a friendly URL, Ajax, and SEO. It was implied about some contradictory indications. With the discussion passing on to friendly URLs are great if for no reason than log analysis, however, friendly URLs are supposedly better for SEO also (besides all
the other stuff for SEO).
Suppose if you would like to start moving over to AJAX for CMS-related, stuff, such as loading a news article when the user clicks on a headline, and stil make the site really search engine friendly.
The added benefit to backloading your ajaxification is that regardless of js enablement, the user could right click your link and open in new tab or window or even bookmark it and it’s still a plain link that will open normally. (Mark Armendariz at NYPHP).
The overall suggestion is to use prototype.js, and some implementation to use asynchronous fetch.