Lightning fast - Using Varnish to accelerate your web applications - Varnish Summit London 2017

Slides for my presentation at the Varnish Summit 2017 in London.

More details: https://talks.feryn.eu/talks/166/lightning-fast-using-varnish-to-accelerate-your-web-applications-varnish-summit-london

1. Lightning fast Using Varnish to accelerate your web applications By Thijs Feryn
2. Slow websites suck
3. Web performance is an essential part of the user experience
4. Down Slowdown ~ downtime
5. Hi, I’m Thijs
6. I’m @ThijsFeryn on Twitter
7. I’m an Evangelist At
9. Mo money Mo servers
10. Write better code
11. Cache
12. Don’t recompute if the data hasn’t changed
14. How does Varnish work?
15. Normally User Server
16. With Varnish User Varnish Server
18. Why is it so good?
19. Smart kernel tricks
20. Stores HTTP output in memory*
21. Respects HTTP best practices
22. Varnish Configuration Language
23. What about TLS/SSL?
24. Quick install & configure
25. apt-get install apt-transport-https curl https://repo.varnish-cache.org/GPG-key.txt | apt-key add - echo "deb https://repo.varnish-cache.org/debian/ jessie varnish-4.1"\ >> /etc/apt/sources.list.d/varnish-cache.list apt-get update apt-get install varnish Install on Linux (Debian) There's stuff for Ubuntu, RHEL & CentOS too
26. -a :80 \ -a :81,PROXY \ -T localhost:6082 \ -f /etc/varnish/default.vcl \ -S /etc/varnish/secret \ -s malloc,3g \ -j unix,user=myUser Simple config
27. Varnish speaks HTTP
28. ✓Idempotence ✓State ✓Expiration ✓Conditional requests ✓Cache variations Varnish speaks HTTP
29. ✓GET ✓HEAD -POST -PUT -PATCH -DELETE -TRACE -OPTIONS Idempotency Only cache GET & HEAD
30. ✓Doesn't cache when cookies are present ✓Doesn't cache when cookies are set ✓Doesn't cache when Autorization headers are present State Avoids caching user- specific content Affects hit rate
31. Expiration Expires: Sat, 09 Sep 2017 14:30:00 GMT Cache-control: public, max-age=3600, s-maxage=86400 Cache-control: private, no-cache, no-store
32. Expiration precedence 1. beresp.ttl (VCL) 2. s-maxage 3. max-age 4. expires 5. 120 seconds by default
33. Conditional requests HTTP/1.1 200 OK Host: localhost Etag: 7c9d70604c6061da9bb9377d3f00eb27 Content-type: text/html; charset=UTF-8 Hello world output GET /if_none_match.php HTTP/1.1 Host: localhost User-Agent: curl/7.48.0
34. Conditional requests HTTP/1.0 304 Not Modified Host: localhost Etag: 7c9d70604c6061da9bb9377d3f00eb27 GET /if_none_match.php HTTP/1.1 Host: localhost User-Agent: curl/7.48.0 If-None-Match: 7c9d70604c6061da9bb9377d3f00eb27
35. Conditional requests HTTP/1.1 200 OK Host: localhost Last-Modified: Fri, 22 Jul 2016 10:11:16 GMT Content-type: text/html; charset=UTF-8 Hello world output GET /if_none_match.php HTTP/1.1 Host: localhost User-Agent: curl/7.48.0
36. Conditional requests HTTP/1.0 304 Not Modified Host: localhost Last-Modified: Fri, 22 Jul 2016 10:11:16 GMT GET /if_none_match.php HTTP/1.1 Host: localhost User-Agent: curl/7.48.0 If-Last-Modified: Fri, 22 Jul 2016 10:11:16 GMT
37. ✓Hostname (or IP) ✓URL Cache variations Identify object in cache
38. Cache variations GET / HTTP/1.1 Host: localhost Accept-Language: nl HTTP/1.1 200 OK Host: localhost Vary: Accept-Language Hallo, deze pagina is in het Nederlands geschreven.
39. Built-in VCL cheat sheet
40. ✓Only GET & HEAD ✓No Cookie ✓No Authorization Built-in VCL (frontend) Lookup in cache Otherwise pass and don't cache
41. ✓No Set-Cookie ✓TTL > 0 ✓No "no-cache", "private", "no-store" Built-in VCL (backend) Store in cache Otherwise blacklist for 120s
42. The actual VCL code
43. sub vcl_recv { if (req.method == "PRI") { /* We do not support SPDY or HTTP/2.0 */ return (synth(405)); } if (req.method != "GET" && req.method != "HEAD" && req.method != "PUT" && req.method != "POST" && req.method != "TRACE" && req.method != "OPTIONS" && req.method != "DELETE") { /* Non-RFC2616 or CONNECT which is weird. */ return (pipe); } if (req.method != "GET" && req.method != "HEAD") { /* We only deal with GET and HEAD by default */ return (pass); } if (req.http.Authorization || req.http.Cookie) { /* Not cacheable by default */ return (pass); } return (hash); }
44. sub vcl_pipe { # By default Connection: close is set on all piped requests, to stop # connection reuse from sending future requests directly to the # (potentially) wrong backend. If you do want this to happen, you can undo # it here. # unset bereq.http.connection; return (pipe); } sub vcl_pass { return (fetch); } sub vcl_hash { hash_data(req.url); if (req.http.host) { hash_data(req.http.host); } else { hash_data(server.ip); } return (lookup); }
45. sub vcl_purge { return (synth(200, "Purged")); } sub vcl_hit { if (obj.ttl >= 0s) { // A pure unadultered hit, deliver it return (deliver); } if (obj.ttl + obj.grace > 0s) { // Object is in grace, deliver it // Automatically triggers a background fetch return (deliver); } // fetch & deliver once we get the result return (miss); } sub vcl_miss { return (fetch); } sub vcl_deliver { return (deliver); }
46. sub vcl_backend_fetch { return (fetch); } sub vcl_backend_response { if (beresp.ttl <= 0s || beresp.http.Set-Cookie || beresp.http.Surrogate-control ~ "no-store" || (!beresp.http.Surrogate-Control && beresp.http.Cache-Control ~ "no-cache|no- store|private") || beresp.http.Vary == "*") { /* * Mark as "Hit-For-Pass" for the next 2 minutes */ set beresp.ttl = 120s; set beresp.uncacheable = true; } return (deliver); }
47. Reality sucks
50. Cache-control ?
51. Legacy
52. Write VCL
53. Normalize
54. vcl 4.0; import std; sub vcl_recv { set req.http.Host = regsub(req.http.Host, ":[0-9]+", ""); set req.url = std.querysort(req.url); if (req.url ~ "\#") { set req.url = regsub(req.url, "\#.*$", ""); } if (req.url ~ "\?$") { set req.url = regsub(req.url, "\?$", ""); } if (req.restarts == 0) { if (req.http.Accept-Encoding) { if (req.http.User-Agent ~ "MSIE 6") { unset req.http.Accept-Encoding; } elsif (req.http.Accept-Encoding ~ "gzip") { set req.http.Accept-Encoding = "gzip"; } elsif (req.http.Accept-Encoding ~ "deflate") { set req.http.Accept-Encoding = "deflate"; } else { unset req.http.Accept-Encoding; } } } }
55. sub vcl_recv { if (req.url ~ "^/status\.php$" || req.url ~ "^/update\.php$" || req.url ~ "^/admin$" || req.url ~ "^/admin/.*$" || req.url ~ "^/user$" || req.url ~ "^/user/.*$" || req.url ~ "^/flag/.*$" || req.url ~ "^.*/ajax/.*$" || req.url ~ "^.*/ahah/.*$") { return (pass); } } URL blacklist
56. sub vcl_recv { if (req.url ~ "^/products/?" return (hash); } } URL whitelist
57. Cookies
58. vcl 4.0; sub vcl_recv { set req.http.Cookie = regsuball(req.http.Cookie, "has_js=[^;]+(; )?", ""); set req.http.Cookie = regsuball(req.http.Cookie, "__utm.=[^;]+(; )?", ""); set req.http.Cookie = regsuball(req.http.Cookie, "_ga=[^;]+(; )?", ""); set req.http.Cookie = regsuball(req.http.Cookie, "_gat=[^;]+(; )?", ""); set req.http.Cookie = regsuball(req.http.Cookie, "utmctr=[^;]+(; )?", ""); set req.http.Cookie = regsuball(req.http.Cookie, "utmcmd.=[^;]+(; )?", ""); set req.http.Cookie = regsuball(req.http.Cookie, "utmccn.=[^;]+(; )?", ""); set req.http.Cookie = regsuball(req.http.Cookie, "__gads=[^;]+(; )?", ""); set req.http.Cookie = regsuball(req.http.Cookie, "__qc.=[^;]+(; )?", ""); set req.http.Cookie = regsuball(req.http.Cookie, "__atuv.=[^;]+(; )?", ""); set req.http.Cookie = regsuball(req.http.Cookie, "^;\s*", ""); if (req.http.cookie ~ "^\s*$") { unset req.http.cookie; } } Remove tracking cookies
59. vcl 4.0; sub vcl_recv { if (req.http.Cookie) { set req.http.Cookie = ";" + req.http.Cookie; set req.http.Cookie = regsuball(req.http.Cookie, "; +", ";"); set req.http.Cookie = regsuball(req.http.Cookie, ";(PHPSESSID)=", "; \1="); set req.http.Cookie = regsuball(req.http.Cookie, ";[^ ][^;]*", ""); set req.http.Cookie = regsuball(req.http.Cookie, "^[; ]+|[; ]+$", ""); if (req.http.cookie ~ "^\s*$") { unset req.http.cookie; } } } Only keep session cookie
60. sub vcl_recv { if (req.http.Cookie) { set req.http.Cookie = ";" + req.http.Cookie; set req.http.Cookie = regsuball(req.http.Cookie, "; +", ";"); set req.http.Cookie = regsuball(req.http.Cookie, ";(language)=", "; \1="); set req.http.Cookie = regsuball(req.http.Cookie, ";[^ ][^;]*", ""); set req.http.Cookie = regsuball(req.http.Cookie, "^[; ]+|[; ]+$", ""); if (req.http.cookie ~ "^\s*$") { unset req.http.cookie; return(pass); } return(hash); } } sub vcl_hash { hash_data(regsub( req.http.Cookie, "^.*language=([^;]*);*.*$", "\1" )); } Language cookie cache variation
61. Alternative language cache variation
62. sub vcl_hash { hash_data(req.http.Accept-Language); } HTTP/1.1 200 OK Host: localhost Vary: Accept-Language Hash language Or just use
63. Cookie guidelines
64. ✓Don't throw everything in a session ✓Use dedicated cookies ✓Strip off tracking cookies in VCL ✓Match cookies in VCL & make decisions ✓Perform cache variations on cookie values ✓Only use session cookies when you really have to ✓Use JWT to store session state client-side Cookie guidelines
65. Block caching
67. Code renders single HTTP response
68. Lowest denominator: no cache
69. <esi:include src="/header" /> Edge Side Includes ✓Placeholder ✓Parsed by Varnish ✓Output is a composition of blocks ✓State per block ✓TTL per block
70. sub vcl_recv { set req.http.Surrogate-Capability = "key=ESI/1.0"; } sub vcl_backend_response { if (beresp.http.Surrogate-Control ~ "ESI/1.0") { unset beresp.http.Surrogate-Control; set beresp.do_esi = true; } } Edge Side Includes
71. ESI vs AJAX
72. <esi:include src="/header" /> Choose wisely <hx:include src="/header"></hx:include> ESI, parsed by Varnish HInclude, parsed by Javascript
73. Breaking news isn't breaking
74. Purging
75. acl purge { "localhost"; "127.0.0.1"; "::1"; } sub vcl_recv { if (req.method == "PURGE") { if (!client.ip ~ purge) { return (synth(405, “Not allowed.”)); } return (purge); } } Purging
76. acl purge { "localhost"; "127.0.0.1"; "::1"; } sub vcl_backend_response { set beresp.http.x-url = bereq.url; set beresp.http.x-host = bereq.http.host; } sub vcl_deliver { unset resp.http.x-url; unset resp.http.x-host; } sub vcl_recv { if (req.method == "PURGE") { if (!client.ip ~ purge) { return (synth(405, "Not allowed")); } if(req.http.x-purge-regex) { ban("obj.http.x-host == " + req.http.host + " && obj.http.x-url ~ " + req.http.x-purge-regex); } else { ban("obj.http.x-host == " + req.http.host + " && obj.http.x-url == " + req.url); } return (synth(200, "Purged")); } } Banning
77. Banning curl -XPURGE "http://example.com/products" curl -XPURGE -H "x-purge-regex:/products" \ "http://example.com" varnishadm> ban obj.http.x-host == example.com && obj.http.x-url ~ ^/product/[0-9]+/details
78. Want more?
79. VMODs
81. Write less VCL Write more code
82. https://github.com/ThijsFeryn/ cacheable-site-silex/tree/v2
85. https://twitter.com/thijsferyn https://instagram.com/thijsferyn https://blog.feryn.eu https://talks.feryn.eu https://book.feryn.eu https://youtube.com/thijsferyn https://soundcloud.com/thijsferyn http://itunes.feryn.eu