Tuning Your Engine
- 1. Tuning Your Engine Joel Bradbury
- 2. I’m Joel Bradbury @joelbradbury View Creative I’m the lead developer at viewcreative.co.uk
- 3. What we can tune
- 6. The Server (Apache etc.) The tuning options
- 7. The Server Extra magic dust (Apache etc.) (squid cache, load balancers etc.) The tuning options
- 8. The Server Your Site Extra magic dust (Apache etc.) (EE for us) (squid cache, load balancers etc.) The tuning options
- 9. The Server Your Site Extra magic dust (Apache etc.) (EE for us) (squid cache, load balancers etc.) The tuning options
- 11. The html The tuning options
- 12. The html The EE & CI Core The tuning options
- 13. The html OUR The EE implementation & CI Core The tuning options
- 14. The html OUR The EE implementation & CI Core The tuning options
- 15. Performance
- 17. Performance •Why it matters •Why it’s an issue
- 18. Performance •Why it matters •Why it’s an issue •What we can do about it
- 19. Why it matters A user’s overall site experience will be set by the worst part
- 20. Why it’s an issue
- 21. Why it’s an issue especially for ExpressionEngine
- 22. Flexibility
- 24. What we can do about it
- 25. What we can do about it Tune it to hell and back
- 26. What can we tune?
- 27. What can we tune? •Site Baseline - disable what you don’t need
- 28. What can we tune? •Site Baseline - disable what you don’t need •Data Structure - you know your data best
- 29. What can we tune? •Site Baseline - disable what you don’t need •Data Structure - you know your data best •Addons - the black boxes in the system
- 30. What can we tune? •Site Baseline - disable what you don’t need •Data Structure - you know your data best •Addons - the black boxes in the system •Templates - the devil is in the details
- 31. Metrics If we’re going to tune it, we’ve got to be able to measure it
- 32. Turn them on Admin > System Administration > Output and Debugging
- 33. Output Profiler
- 34. Output Profiler
- 35. Output Profiler
- 37. Template Logging
- 38. Template Logging
- 39. Template Graph
- 40. Time Template Graph
- 41. Time Template execution Template Graph
- 42. End Metrics
- 43. End Metrics 15 Database queries
- 44. End Metrics 15 Database queries Execution Profile
- 45. End Metrics 15 Database queries Execution Profile * take all numbers/times as guides only. yours will differ
- 49. Clean Template
- 51. Clean Template 6-8 Queries Execution profile
- 61. (0.008751) Template Parsing Finished Keeping it in context
- 64. Templates
- 66. Loops can kill
- 67. Loops can kill your performance
- 68. Loops can kill your performance but only if you’re sloppy
- 69. The Cost of a Loop using channel:entries as an example {exp:channel:entries dynamic=“no” channel=“example” limit=“1”} ...
- 70. The Cost of a Loop using channel:entries as an example clean template
- 71. The Cost of a Loop using channel:entries as an example loop template
- 72. The Cost of a Loop using channel:entries as an example loop template
- 73. The Cost of a Loop using channel:entries as an example
- 74. The Cost of a Loop using channel:entries as an example
- 75. The Cost of a Loop using channel:entries as an example limit = 100
- 76. The Cost of a Loop using channel:entries as an example limit = 500
- 77. The Cost of a Loop using channel:entries as an example limit = 1000
- 78. The Cost of a Loop using channel:entries as an example limit = 2500
- 79. The Cost of a Loop the effect on the database queries
- 80. The Cost of a Loop the effect on the database queries 24 100 entries
- 81. The Cost of a Loop the effect on the database queries 24 24 100 500 entries entries
- 82. The Cost of a Loop the effect on the database queries 24 24 24 100 500 1000 entries entries entries
- 83. The Cost of a Loop the effect on the database queries 24 24 24 24 100 500 1000 2500 entries entries entries entries
- 84. The Cost of a Loop the effect on the database queries with current db_cache
- 85. The Cost of a Loop the effect on the database queries 6-8 for n entries with current db_cache
- 88. Lather, Repeat, Rinse using an “optimised” channel:entries as an example {exp:channel:entries dynamic=“no” channel=“example” limit=“1” disable=“{disable_entries} ” cache=“yes” refresh=“900”}
- 89. Lather, Repeat, Rinse with everything disabled
- 90. Lather, Repeat, Rinse with everything disabled 21 for n entries
- 91. Lather, Repeat, Rinse with tag caching on uncached
- 92. Lather, Repeat, Rinse with tag caching on caches primed
- 93. What we want:
- 94. What we want: • Caches shouldn’t be relied on to get good performance
- 95. What we want: • Caches shouldn’t be relied on to get good performance • Entry count shouldn’t affect the number of Database queries
- 96. What we want: •Caches shouldn’t be relied on to get good performance •Entry count shouldn’t affect the number of Database queries •Entry count should relate linearly to template processing time
- 97. What we want: •Caches shouldn’t be relied on to get good performance •Entry count shouldn’t affect the number of Database queries •Entry count should relate linearly to template processing time •What works for 10, needs to scale to work for 10,000
- 98. Nested Loops
- 99. Nested Loops “ Pro tip: don't ever, ever nest a channel:entries tag inside another tag that loops. Srsly. Don't. #eecms @low
- 100. Nested Loops “ Pro tip: don't ever, ever nest a channel:entries tag inside another tag that loops. Srsly. Don't. #eecms @low
- 101. Nested Loops Category Archive Page
- 102. Nested Loops Category Archive Page Category Loop {
- 103. Nested Loops Category Archive Page } Entries Loop Category Loop { } Entries Loop } Entries Loop
- 104. Nested Loops Category Archive Page
- 105. Nested Loops Category Archive Page What you’d expect :
- 106. Nested Loops Category Archive Page What you’d expect : +2 for category loop
- 107. Nested Loops Category Archive Page What you’d expect : +2 for category loop +3 for entries loop
- 108. Nested Loops Category Archive Page What you get :
- 109. Nested Loops Category Archive Page What you get : +2 for category loop
- 110. Nested Loops Category Archive Page What you get : +2 for category loop +3
- 111. Nested Loops Category Archive Page What you get : +2 for category loop for EACH entries loop +3 PER CATEGORY
- 112. Nested Loops Category Archive Page With some tuning (and some php) you can get:
- 113. Nested Loops Category Archive Page With some tuning +2 for category loop (and some php) you can get:
- 114. Nested Loops Category Archive Page With some tuning +2 for category loop (and some php) you can get: +3 for entries loop
- 115. Nested Loops Category Archive Page With some tuning +2 for category loop (and some php) you can get: +3 for entries loop see : bit.ly/eeci2010_low
- 116. Where the pain comes from
- 117. Where the pain comes from Never testing with end-scale datasets
- 118. Key Points
- 119. Key Points
- 120. Key Points •Don’t rely on caches for performance
- 121. Key Points •Don’t rely on caches for performance •Measure as you go
- 122. Key Points •Don’t rely on caches •Test with end-scale for performance data sets •Measure as you go
- 123. Key Points •Don’t rely on caches •Test with end-scale for performance data sets •Measure as you go •Understand what’s going on
- 124. Thank you Tuning your Engine @joelbradbury @viewcreative
Editor's Notes
- #22: All things to all men Implement the 30% features you need, you could do it quicker
- #24: - Micheal Boiynk addon has the right idea
- #25: - Micheal Boiynk addon has the right idea
- #26: - Micheal Boiynk addon has the right idea
- #27: - Micheal Boiynk addon has the right idea
- #43: From the Agile records home page
- #44: From the Agile records home page
- #45: From the Agile records home page
- #90: Embeds, loops, conditionals etc... (time)
- #118: Need to do this as you go
- #157: A wise old man told me I should end with key points