AutoBlogged Support
posted this on November 29, 2010 15:19
We have been receiving lots of feedback from our customers and we are surprised with the variety of sites people have built with AutoBlogged. Autoblogs have traditionally had a bad name as spammy or as content thieves but some of you have used AutoBlogged to build some very useful web sites.
One thing we are seeing is people using AutoBlogged in ways we really never considered. AutoBlogged works best when you give it two or three RSS feeds based on various searches. However, when you load up hundreds of feeds and have complex filtering requirements, you might notice a significant hit on your site's performance.
Part of the problem is that PHP, as an interpreted language, will always require more overhead than a compiled application. Part of the problem is that WordPress is a complex platform that already does an enormous amount of processing for each page view. Part of it is that AutoBlogged does quite a bit of work to process and tag each post.
While much has been done to optimize the performance of AutoBlogged, how you configure the plugin can have a significant impact on the load it puts on the server, especially when the plugin adds a large number of posts each day.
If you want to get maximum performance from AutoBlogged, here are some things you can do:
1. Remote Filtering - Because all filtering tasks increase script processing load, try to limit your use of feed-level filters or search and replace operations. Filtering requires repetitive searching that could potentially have an impact on a busy site.Try to offload as much filtering as possible on the remote end.
For example, use advanced search options with Google Blog Search to filter out unwanted words, limit to a specific date range, or specify the language. Then under Filtering, clear all the words from the keywords blacklist. Also consider using Yahoo! Pipes, MySyndicaat, or another feed aggregator with filtering capabilities to fine-tune your source feed. Anything you can do to move the processing off your server means that much less work your server has to do.
The feed level filters and search and replace filters are useful for simple processing but can quickly slow down the script if you overuse that feature. If you need more advanced filtering capabilities, we suggest using Yahoo! Pipes, MySyndicaat, or another feed aggregator with filtering capabilities.
2. Limit the Number of Feeds – Although there is no specific limit to the number of feeds AutoBlogged can handle, adding too many feeds can slow down the process and possibly result in script timeouts. Again, an external feed aggregator is an excellent solution.
3. Do not Retrieve the Original Article – AutoBlogged by default will visit the URL of the original article in order to gather additional keywords to use as tags. Skipping this step will save a significant amount of CPU usage, will reduce network traffic, but will limit the effectiveness of the built-in tagging engine. Under Tag Options, only check the box to use original tags from feed. This will save a visit to the original URL and the subsequent parsing of tags. Instead of automatically parsing tags from the original URL, under Tag Options, use the Additional Tags box as a random source for tags for each post.
4. Do not Search for Existing Categories – AutoBlogged has two options for dealing with existing blog categories that appear in an article: it can add that category to the post or it can add it as a tag. This is very useful for automatically categorizing each post but it also means that AutoBlogged must loop through each blog category to see if it exists in the post.
5. Limit Duplicate Matching – To prevent duplicate posts from appearing AutoBlogged will search for duplicates based on the post title or based on the original link. Filtering by title works best in some situations but filtering by link works better in others. We do not recommend using both at the same time because AutoBlogged must perform a database lookup for each one.
6. Limit Your Plugins – If you find that WordPress in general is slow, you should take a look at your plugins and consider only enabling the bare minimum. Remember that most of these plugins will run with every page load.
7. Limit the Features you Use – If you have a very busy blog and limited CPU resources, you may have to limit the AutoBlogged features you use. AutoBlogged can do quite a bit but sometimes you may want it not do so much to help performance. Image and video processing, checking to see if links already exist in your blog, checking to see if the author exists, and saving images locally all require extra processing that can slow things down. At some point you need to decide which is most important: features, performance, or the amount of money you spend on server equipment.
Comments
Thanks for a great plugin and for the great support.
I was reading the above and have been amending our settings accordingly. I also remember reading that timeout settings and PHP settings can affect the performance of autoblogged and its capacity to ingest feeds. We run our own server and can max it out to make autoblogged happy, what settings would you advise amending on the server to increase the number of feeds autoblogged can ingest? I only ask this as we run our own server.
Many thanks in advance.
Regards
Mike