Options and Settings of the Component
This page as mostly a demonstration purpose. Please visit Administrator Component for more information
Checking
Static path prefixes
These paths define the locations of static resources. On most Joomla websites, that will be the image's and in fewer cases the templates/ folder.
How many links to check in one cron batch.
Compared to extracting links, checking them is far more time-consuming.
You can set different batch sizes for the HTTP and the Command Line Interface (CLI) crons, as the latter has less of a time restriction.
Typically, a website running PHP has a timeout (max_execution_time) of 30 - 60 seconds. The default is 30 seconds. The CLI usually has no time limit.
The max_execution_time divided by the Request Timeout setting gives a rough upper limit for the amount of links that could be checked in one batch. You will find an estimate on the setup page.
Both the extraction as the checker phase are build as Failed? No Sweat. Let's try again latter. So it wouldn't matter if your job reaches the PHP timeout time. But it is cleaner to stay within it.
The administrator Pseudo cron always checks one link per request.
Maximum time to wait for a response when checking a link.
You can set different values for the HTTP and the CLI cron, as the latter as less of a time restriction. The administrator Pseudo cron uses the HTTP setting
There is a fine line between broken, maybe broken and working. With this option, the checker marks links that appear broken but most likely work for a normal visitor as a warning.
(This option is most likely removed in a future version)
This option is intended to reduce the amount of data transferred. With this option On the checker sends a HEAD (are you there) request to the remote server. Most servers response correctly with static content like images and stylesheets.
If the response is any kind of error or redirect, thus a response other than 200 the link is rechecked as GET.
Recheck link after is the recheck interval for all links, broken or working. Links marked to ignore are always skipped.
With the Recheck broken link after setting, you can set a shorter interval for broken links, to catch temporality failures.
This option is intended to reduce the amount of data transferred. With this option On the checker request only a part of the remote content. Successful Range request have a response code 206.
Usually, Range request only works with static content like images and zip files.
If the response is any kind of error or redirect, thus a response other than 200 or 206, the link is rechecked using a normal GET.
So with Use Head request and Use Range request the request are:
- HEAD
- response 200, perfect finish checking
- GET with range header
- response 200 or 206, perfect finish checking
- GET without range header
How often should broken links be rechecked.
The checker will always try to get the final destination of a link. This allows for easy replacement of the links.
It either uses CURLOPT_FOLLOWLOCATION or checks a link recursively bases on a redirect response code (301, 302, 303, 307) and the Location header.
If your server as allow_url_open restrictions, redirects can not be followed directly by curl, so the recursive method is used.
If you want to disable following redirects completely, disable this option.
The 'Max Redirects' number controls the maximum number of hops to follow. If this number is reached the link is reported as broken with custom response 'Too many redirects'
Depending on the response, the checker can keep the response from the remote server in the log.
Never
Disable saving any response, reducing database size.
Auto
Log the most common text responses if avaiable.
If the server responds with a 200 code on a HEAD Request, no content is exchanged and therefor nothing is logged.
If the server responds with a 206 - Partial content on a RANGE Request, only a part of the response is sent and therefor logged.
Always
Keep all the responses. This disables the range and head options
Text
Keep all text like responses. This disables the range and head options
Some plugins, like the External Link Extractor, will override this setting when needed.
A website might contain links that can not be checked, like mailto: links.
With this option, you can either show them in the reports (visible with the unchecked protocol filter) or ignore them completely.
Some websites serve content based on the browsers language
This tells the remote server which language you prefer. The automatic setting is based on your website's default language.
These options are used to masquerade the checker as a browser. Some Web Application firewalls (WAFs) respond friendlier if Cookies are set and match the User-Agent.
You can create one or more custom signature be adding a json file to [ROOT]/administrator/com_blc/forms/signatures.
Example:
{
"userAgent": "Some ser agent",
"headers": [
"Accept-Encoding: deflate, br",
"Sec-Fetch-Dest: document",
"Sec-Fetch-User: ?1",
"Sec-Fetch-Mode: navigate",
"Sec-Fetch-Site: none",
],
"Accept-Language": "Language string / Optional"
}The Accept-Language is optional, if set it will override the setting from the BLC configuration
With this option enabled, the Fetch metadata request header are adjusted according to the URL location.
Sec-Fetch-Site
This option is set to same-site for internal links and cross-site for external links.
Sec-Fetch-Mode
Should be set in the signature to 'navigate' as if a user clicks on a link to open a html-page, image, video or whatever.
Sec-Fetch-Dest
Should be set in the signature to 'document', as if a user views the destination in the browser.
The checker does not distinguish between links in an <img>-tag and <a>-tags.
With this option On , SSL certificates of remote websites are checked. Websites with invalid certificates will be reported as broken with a pseudo response SSL Certificate Error (606).
Depending on your servers (mis)configuration, SSL validation might fail for al websites or cause timeouts.
Invalid Server certificates
The most common issues are with SSL certificates for the domain itself:
- Expired
- Invalid domain
- SSL certificate problem: unable to get local issuer certificate
These problems should be visible in your web browser as well. Hopefully the owner of the website will resolve the issue soon.
SSL Chain problems
The SSL certificate chain is the list of certificates that contains the SSL certificate, intermediate certificate authorities, and root certificate authority that enables the connecting device to verify that the SSL certificate is trustworthy. A server should send the intermediate certificate(s). However, some servers are misconfigured and only send the SSL certificate itself. Web browsers are quite forgiving, a tool like curl is not.
You can check a certificate on a website like: https://www.ssllabs.com/ssltest/analyze.html. Message like `This server's certificate chain is incomplete`, `Chain issues Incomplete`, `SSL certificate problem: unable to get local issuer certificate` and `NOT TRUSTED` indicate some SSL chain issue.
It is possible to solve these issues on the server. This is quite technical and requires root access.
As SSL connection errors are not reported as HTTP Response codes, the checker uses custom codes to report SSL problems
It seems that some WAF's block browsers polling for different TLS versions. They expect Chrome to use TLS v1.3 without bothering about the older versions.
This option enforces the checker to use the selected version only.
Most modern web servers should support TLS v1.3, so setting this option to this newest version should work fine for most.
The checker will encounter older servers where the SSL Connection fails (CURLE_SSL_CONNECT_ERROR (35)). In that case the checker will retry using TLS v1.2.
- default: use TLS v1.2 or TLS v1.3
- TLS 1.2: use TLS v1.2 and if that fails, fallback to default
- TLS 1.3: use TLS v1.3 and if that fails, fallback to default.
This is an advanced feature best left to default.
A certificate authority (CA) is a trusted entity that issues Secure Sockets Layer (SSL) certificates for websites. Like a kind of signature to ensure the web server's certificate is valid and issued to the owner of the domain.
If you are getting a lot of SSL Chain Errors your server might have an outdated certificate authority (CA) bundle. In that case you could try the Certificate Bundle shipped with Joomla (Bundled), try a search on the System or provide your own bundle.
This option sets the CURL option CURLOPT_CAINFO, in addition CURL used the CURLOPT_CAPATH option to find certificates. So this option can be used to add missing intermediate certificates.
This option has no effect on CURL implementations based on Schannel library (windows).
This enabled verbose logging (CURLOPT_VERBOSE).
This adds information about the SSL -handshake, request and response headers to the Log.
Effective after the link is checked with this option On.
This limits the amount of request to your own or an external website.
Be friendly to your neighbors!
Links are only (rec)checked if the throttle time has expired. The HTTP and Scheduded Task action will skip the throttled domain and continue with other links, fetching new links from the database until Checks per HTTP batch is reached, or all links are considered.
The admin pseudo cron will simply skip the link. If this bothers, set the value for Pause between checks to a value larger than Internal/External Domain throttle and you will never hit a throttled domain.
For the CLI you can either skip a throttled domain or pause checking using the option Handle throttle with CLI/ If you opt to skip links, the actual number of checked links might be lower than the configured Checks per CLI batch.
This is a list of hosts that should never be checked.
For example, Facebook has a login restriction on a lot of pages.
You can either ignore these links completely or report them marked to ignore using the What to do with these domains option
Pattern matching on the path of links (https://example.com/path/).
Intended to exclude internal links, but should work on all links.
As with the Domains above, you can control the reporting with What to do with the matches
The field is intended for domain(s) that always redirect, for example:
- Links to a URL shortener service
- Affiliate links
- Google services like drive and maps.
The redirects are followed to the final destination, so the checker should still report broken links (400+ Response codes) but will not show any redirects. That includes 'valid' redirects at the final destination.
Valid links will have a Redirect Ignored custom response code.
Mail Reports
With this option enabled the report will include all items (articles, categories etc) where the link was found. Including the anchor and links to view and edit the item.
The minimum interval between two reports.
How many links should be reported. This is per 'what'
If set to yes the report will only be sent if there are new (broken) links.
If an existing link appears on a new location, a link is still considered old.
With the options: Broken, Warning, Redirect, Parked and New you can select what types of (broken) links you want to report.
Run the report after a http of cli extract operation.
Whether a report is sent still depends on the frequency.
This option is not active when using the Admin pseudo cron.
Run the report after a http of cli check operation.
Whether a report is sent still depends on the frequency.
This option is not active when using the Admin pseudo cron.
One or more Joomla Users to receive the report.