copyright | lastupdated | keywords | subcollection | ||
---|---|---|---|---|---|
|
2024-10-09 |
Use Page Rules, standard cache levels, Custom Caching Sets |
cis |
{{site.data.keyword.attribute-definition-list}}
{: #use-page-rules-with-caching}
Page rules give you the ability to take various actions based on the page's URL, such as creating redirects, fine-tuning caching behavior, or enabling and disabling services. {: shortdesc}
A page rule takes effect on a URL pattern that matches the following format:
<scheme>://<hostname><:port>/<path>
For example,
https://www.example.com:80/image.png
The scheme
and port
components are optional. If the scheme
component is omitted, the format accepts http://
and https://
protocols. If the port
is not specified, the rule matches all ports. You can perform basic wildcard matches by using a *
symbol in your rule pattern, allowing it to match a series of similar patterns.
{: #page-rule-considerations}
- Only one page rule takes effect on any request.
- Page rules get priority in an order from top to bottom. After a URL matches a rule, only that rule is applied. In other words, if a page rule triggered already on a request, any subsequent rules that also match the URL pattern do not take effect.
- Generally, it is recommended to order from most-specific to least-specific rules.
- Page rules can be disabled, in which case they take no action. However, you can still see the rules in the list and edit them. Setting the Enabled toggle to Off creates a page rule that is disabled initially.
{: #forwarding-url-redirection}
Redirect one URL to another using an HTTP 301 or 302 redirect. The contents of any section of a URL that a wildcard matches can be referenced using $X
syntax. The X
indicates the index of a glob in the pattern: $1
is replaced with the first wildcard match, $2
with the second wildcard match, and so on.
For example, suppose you set the following rule where:
- URL match =
*.example.com/*
- Enabled = On
- Setting = Forwarding URL
- Set as = 301 Permanent Redirect to
http://example.com/$2/
Here, a request to www.example.com/stuff/things
is redirected to http://example.com/stuff/things
.
Be careful not to create a redirect in which the domain points to itself as a destination. This mistake can cause an infinite redirect error, and the affected URLs do not resolve. {: note}
{: #redirecting-to-https}
If you want to redirect your visitors to use HTTPS, set the rule behavior to the Always Use HTTPS setting instead.
{: #custom-caching}
Set caching behavior for any URL matching the page rule pattern, using any of the standard cache levels. Setting Cache Level to Cache Everything caches any content, even if it is not one of the default static file types. Setting Cache Level to the Bypass setting prevents caching on that URL.
When you specify a cache level by using page rules, you can set an Edge Cache TTL, which controls how long CIS retains files in the cache.
Browser Cache TTL controls how long resources that are cached by client browsers remain valid. If a browser requests a resource again and the TTL isn't expired, the browser receives an HTTP 304 (Not Modified)
response. You can set TTLs ranging from 30 minutes to 1 year.
Not all default caching behaviors are strictly RFC-compliant. Setting Origin Cache Control through page rules uses a newer set of caching rules that seeks to adhere more closely to RFCs, primarily regarding revalidation. For example, our default behavior with max-age=0
is to not cache at all, whereas setting Origin Cache Control caches, but it always revalidates.
The following example sets a page rule to cache everything found in the /images
folder. Cached resources expire in 30 minutes in the user's browser, and they expire after one day in the IBM CIS data centers.
- URL match =
*.example.com/images/*
- Enabled = On
- Rule behavior settings:
Setting | Set as |
---|---|
Cache level | Cache everything |
Browser Cache TTL | 30 minutes |
Edge Cache TTL | 1 day |
{: caption="Rule behavior settings to cache everything in /images" caption-side="bottom"} |
Serve Stale Content serves pages from our cache, even when your server goes down. Visitors see a limited version of your site, with a message that they are in offline browsing mode.
This feature returns an HTTP status 503. When servers are online again, CIS seamlessly takes visitors to regular browsing.
If the requested page is not in the cache, the visitor sees an error page that informs them the page they are requesting is offline. If a Cache Everything page rule is enabled with expiration times set lower than the caching frequency, the Serve Stale Content is purged in the corresponding interval. {: note}