Content composition with Edge Side Includes

Varnish can create web pages by assembling different pages, called fragments, together into one page. These fragments can have individual cache policies. If you have a web site with a list showing the five most popular articles on your site, this list can probably be cached as a fragment and included in all the other pages.

Used properly this strategy can dramatically increase your hit rate and reduce the load on your servers.

In Varnish we’ve only implemented a small subset of ESI, because most of the rest of the ESI specifications facilities are easier and better done with VCL:

esi:include
esi:remove
<!--esi ...-->

Content substitution based on variables and cookies is not implemented.

Varnish will not process ESI instructions in HTML comments.

Example: esi:include

Lets see an example how this could be used. This simple cgi script outputs the date:

#!/bin/sh

echo 'Content-type: text/html'
echo ''
date "+%Y-%m-%d %H:%M"

Now, lets have an HTML file that has an ESI include statement:

<HTML>
<BODY>
The time is: <esi:include src="/cgi-bin/date.cgi"/>
at this very moment.
</BODY>
</HTML>

For ESI to work you need to activate ESI processing in VCL, like this:

sub vcl_backend_response {
    if (bereq.url == "/test.html") {
       set beresp.do_esi = true; // Do ESI processing
       set beresp.ttl = 24 h;    // Sets the TTL on the HTML above
    } elseif (bereq.url == "/cgi-bin/date.cgi") {
       set beresp.ttl = 1m;      // Sets a one minute TTL on
                                 // the included object
    }
}

Note that set beresp.do_esi = true; is not required, and should be avoided, for the included fragments, unless they also contains <ESI::include …/> instructions.

Example: esi:remove and <!–esi … –>

The <esi:remove> and <!–esi … –> constructs can be used to present appropriate content whether or not ESI is available, for example you can include content when ESI is available or link to it when it is not. ESI processors will remove the start (“<!–esi”) and the end (”–>”) when the page is processed, while still processing the contents. If the page is not processed, it will remain intact, becoming a HTML/XML comment tag. ESI processors will remove <esi:remove> tags and all content contained in them, allowing you to only render the content when the page is not being ESI-processed. For example:

<esi:remove>
  <a href="http://www.example.com/LICENSE">The license</a>
</esi:remove>
<!--esi
<p>The full text of the license:</p>
<esi:include src="http://example.com/LICENSE" />
-->

What happens when it fails ?

By default, the fragments must have resp.status 200 or 204 or their delivery will be considered failed.

Likewise, if the fragment is a streaming fetch, and that fetch fails, the fragment delivery is considered failed.

If you include synthetic fragments, that is fragments created in vcl_backend_error{} or vcl_synth{}, you must set (be)resp.status to 200 before return(deliver);, for example with a return (synth(200)) or return (error(200)) transition.

Failure to properly deliver an ESI fragment has no effect on its parent request delivery by default. The parent request can include the ESI fragment with an onerror attribute:

<ESI:include src="…" onerror="continue"/>

This attribute is ignored by default. In fact, Varnish will treat failures to deliver ESI fragments as if there was the attribute onerror="continue". In the absence of this attribute with this specific value, Varnish should normally abort the delivery of the parent request.

We say “abort” rather than “fail”, because by the time Varnish starts inserting the fragments, the HTTP response header has long since been sent, and it is no longer possible to change the parent requests’s resp.status to a 5xx, so the only way to signal that something is amiss, is to close the connection in the HTTP/1 case or reset the stream for h2 sessions.

However, it is possible to allow individual <ESI:include… to continue in case of failures, by setting:

param.set feature +esi_include_onerror

Once this feature flag is enabled, a delivery failure can only continue if an onerror attribute said so. The ESI specification states that in that case the failing fragment is not delivered, which is honored based on the status code, or based on the response body only when streaming is disabled (see beresp.do_stream).

Can an ESI fragment also use ESI-includes ?

Yes, but the depth is limited by the max_esi_depth parameter in order to prevent infinite recursion.

Doing ESI on JSON and other non-XML’ish content

Varnish will peek at the first byte of an object and if it is not a “<” Varnish assumes you didn’t really mean to ESI process it. You can disable this check by:

param.set feature +esi_disable_xml_check

Ignoring BOM in ESI objects

If you backend spits out a Unicode Byte-Order-Mark as the first bytes of the response, the “<” check will fail unless you set:

param.set feature +esi_remove_bom

ESI on invalid XML

The ESI parser expects the XML to be reasonably well formed, but this may fail if you are ESI including non-XML files. You can make the ESI parser disregard anything but ESI tags by setting:

param.set feature +esi_ignore_other_elements

ESI includes with HTTPS protocol

If ESI:include tags specify HTTPS protocol, it will be ignored by default, because Varnish has no way to fetch it with encryption. If you want Varnish to fetch them like it does anything else, set:

param.set feature +esi_ignore_https

ESI on partial responses (206)

Varnish supports range requests, but in general partial responses make no sense in an ESI context.

If you really know what you are doing, change the 206 to a 200:

sub vcl_backend_response {
    if (beresp.status == 206 && beresp.http.secret == "swordfish") {
        set beresp.do_esi = True;
        set beresp.status = 200;
    }
}

ESI and return(vcl(…))

If the original client request switched to a different VCL using return(vcl(...)) in vcl_recv, any esi:include-requests will still start out in the same VCL as the original did, not in the one it switched to.

ESI and gzip compression

Varnish’s ESI implementation handles gzip compression automatically, no matter how it is mixed: The parent request can be compressed or uncompressed and the fragments can be compressed or uncompressed, it all works out.

Varnish does this compressing all parts of ESI responses separately, and stitching them together on the fly during delivery, which has a negative impact on compression ratio.

When you set beresp.do_esi = True; on a gzip’ed response, it will be uncompressed and recompressed part-wise during the fetch.

The part-wise compression reduces the opportunities for removing redundancy, because back-references in the gzip data stream cannot point outside it’s own part.

The other case where compression ratio is impacted, is if an uncompressed fragment is inserted into a compressed response.