Caching 2020 queries

[raghuramakrishnan71]

Queries for the Caching chapter. Progress on #917
Carried forward the open list to #1498
Todo query list

• Usage of Last-Modified/If-Modified-Since and ETag/If-None-Match compared to the previous year's stats (are they being used more, which is used more often)?
• Whether making a valid conditional request (passing an If-Modified-Since or If-None-Match request header which matches the previous server response header) actually gets a 304 instead of a 200?
• Number of websites using AppCache versus service worker.
• Number of websites using AppCache versus service worker usage by HTTP/HTTPS.
• Cacheable vs non-cacheable responses.
• Distribution of TTL in cacheable responses.
• Desktop cache TTL percentiles by resource type.
• Distribution of cacheability by content type for desktop.
• Distribution of cacheability by content type for mobile.
• Count, percentage of short TTLs
• Median difference between content age and TTL
• Percent of 1st party requests with short TTLs.
• Percent of 3rd party requests with short TTLs.
• Distribution of Lighthouse scores for the "Uses Long Cache TTL" audit for mobile web pages.

Todo document list

• Upto Setting cookies on cacheable responses
• Service workers
• What type of content are we caching (part 1)?
• What type of content are we caching (part 2)?
• What type of content are we caching (part 3)?
• How do cache TTLs compare to resource age (part 1)?

[raghuramakrishnan71]

@HTTPArchive/analysts have added the first sql file, request review (percentiles of TTL by resource type, client - desktop or mobile).

[tunetheweb]

Query itself is fine. Added a couple of comments on convention that you should follow.

Let us know if any of that is not clear.

[rviscomi]

Initial query LGTM.

[raghuramakrishnan71]

@HTTPArchive/analysts request review of ttl_by_resource_third_party, resource_without_cache, and resource_without_cache_third_party.

[tunetheweb]

Couple of comments but looks like queries are good so no need to submit in smaller batches and just request a re-review when rest are done.

[rviscomi]

@raghuramakrishnan71 is this everything? If so I'll give this a thorough review. Otherwise ping this PR when you've added all of the queries.

[rviscomi]

📟 ping @raghuramakrishnan71

[raghuramakrishnan71]

Updated the SQLs after validating with Chapter contents. @HTTPArchive/analysts

[raghuramakrishnan71]

Results saved to the Caching sheet.
https://docs.google.com/spreadsheets/d/1fYmpSN3diOiFrscS75NsjfsrKXzxxhUMNcYSqXnQJQU/edit?ts=5fa1ab38#gid=1338572274

[raghuramakrishnan71]

(rviscomi: moved TODO list to the top comment for easier tracking)

[AbbyTsai]

hello, saw some of our metrics may be similar enough to combine as one. here is reference items, and like to hear your thoughts if you get other ideas.
(19_02) 20_Distribution of potential byte savings from the Lighthouse caching audit.
(19_04) 20_Distribution of Lighthouse scores for the "Uses Long Cache TTL" audit for mobile web pages.

[rviscomi]

@raghuramakrishnan71 getting closer. Let's resolve these last few remaining comments and merge and iterate as needed.

[AbbyTsai]

these look well to me and found a passion inside.

[raghuramakrishnan71]

hello, saw some of our metrics may be similar enough to combine as one. here is reference items, and like to hear your thoughts if you get other ideas.
(19_02) 20_Distribution of potential byte savings from the Lighthouse caching audit.
(19_04) 20_Distribution of Lighthouse scores for the "Uses Long Cache TTL" audit for mobile web pages.

@AbbyTsai - i checked out the compression queries, which are using $.audits.uses-text-compression.score, in the main branch. I am trying to explore if there is a caching specific score (maybe uses-long-cache-ttl) and savings in bytes. Agree, the query may look similar to 19_02, 19_04 as the report field of the lighthouse table is already being used.

[rviscomi]

Merging this to keep it moving. @raghuramakrishnan71 please open up a new PR if there are any more follow up changes needed.