Merge branch 'main' of github.com:HTTPArchive/almanac.httparchive.org…

… into production

.github/workflows/add-to-release-notes.yml

@@ -16,6 +16,6 @@ jobs:
     if: github.repository == 'HTTPArchive/almanac.httparchive.org'
     steps:
     - name: Update release notes
-      uses: release-drafter/[email protected].0
+      uses: release-drafter/[email protected].1
       env:
         GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}

sql/2022/performance/bfcache_cachecontrol_nostore.sql

@@ -0,0 +1,37 @@
+CREATE TEMP FUNCTION HAS_NO_STORE_DIRECTIVE(cache_control STRING) RETURNS BOOL AS (
+  REGEXP_CONTAINS(cache_control, r'(?i)\bno-store\b')
+);
+
+WITH requests AS (
+  SELECT
+    client,
+    rank,
+    LOGICAL_OR(HAS_NO_STORE_DIRECTIVE(JSON_VALUE(payload, '$._cacheControl'))) AS includes_ccns
+  FROM
+    `httparchive.almanac.summary_response_bodies`
+  WHERE
+    date = '2022-06-01' AND
+    firstHtml
+  GROUP BY
+    client,
+    rank,
+    page
+)
+
+SELECT
+  client,
+  _rank AS rank,
+  COUNTIF(includes_ccns) AS pages,
+  COUNT(0) AS total,
+  COUNTIF(includes_ccns) / COUNT(0) AS pct
+FROM
+  requests,
+  UNNEST([1000, 10000, 100000, 1000000, 10000000]) AS _rank
+WHERE
+  rank <= _rank
+GROUP BY
+  client,
+  rank
+ORDER BY
+  rank,
+  client

sql/2022/performance/bfcache_unload.sql

@@ -0,0 +1,40 @@
+WITH lh AS (
+  SELECT
+    _TABLE_SUFFIX AS client,
+    url AS page,
+    JSON_VALUE(report, '$.audits.no-unload-listeners.score') = '0' AS has_unload
+  FROM
+    `httparchive.lighthouse.2022_06_01_*`
+),
+
+pages AS (
+  SELECT
+    _TABLE_SUFFIX AS client,
+    url AS page,
+    rank
+  FROM
+    `httparchive.summary_pages.2022_06_01_*`
+)
+
+
+SELECT
+  client,
+  _rank AS rank,
+  COUNTIF(has_unload) AS pages,
+  COUNT(0) AS total,
+  COUNTIF(has_unload) / COUNT(0) AS pct
+FROM
+  lh
+JOIN
+  pages
+USING
+  (client, page),
+  UNNEST([1000, 10000, 100000, 1000000, 10000000]) AS _rank
+WHERE
+  rank <= _rank
+GROUP BY
+  client,
+  rank
+ORDER BY
+  rank,
+  client

sql/2022/performance/cls_animations.sql

@@ -0,0 +1,25 @@
+WITH lh AS (
+  SELECT
+    _TABLE_SUFFIX AS client,
+    ARRAY_LENGTH(JSON_QUERY_ARRAY(report, '$.audits.non-composited-animations.details.items')) AS num_animations
+  FROM
+    `httparchive.lighthouse.2022_06_01_*`
+)
+
+
+SELECT
+  percentile,
+  client,
+  APPROX_QUANTILES(num_animations, 1000)[OFFSET(percentile * 10)] AS num_animations,
+  COUNTIF(num_animations > 0) AS pages,
+  COUNT(0) AS total,
+  COUNTIF(num_animations > 0) / COUNT(0) AS pct
+FROM
+  lh,
+  UNNEST([10, 25, 50, 75, 90, 100]) AS percentile
+GROUP BY
+  percentile,
+  client
+ORDER BY
+  percentile,
+  client

sql/2022/performance/cls_unsized_image_height.sql

@@ -0,0 +1,24 @@
+WITH lh AS (
+  SELECT
+    _TABLE_SUFFIX AS client,
+    CAST(JSON_VALUE(unsized_image, '$.node.boundingRect.height') AS INT64) AS height
+  FROM
+    `httparchive.lighthouse.2022_06_01_*`,
+    UNNEST(JSON_QUERY_ARRAY(report, '$.audits.unsized-images.details.items')) AS unsized_image
+)
+
+
+SELECT
+  percentile,
+  client,
+  APPROX_QUANTILES(height, 1000)[OFFSET(percentile * 10)] AS height,
+  COUNT(0) AS unsized_images
+FROM
+  lh,
+  UNNEST([10, 25, 50, 75, 90, 100]) AS percentile
+GROUP BY
+  percentile,
+  client
+ORDER BY
+  percentile,
+  client

sql/2022/performance/cls_unsized_images.sql

@@ -0,0 +1,25 @@
+WITH lh AS (
+  SELECT
+    _TABLE_SUFFIX AS client,
+    ARRAY_LENGTH(JSON_QUERY_ARRAY(report, '$.audits.unsized-images.details.items')) AS num_unsized_images
+  FROM
+    `httparchive.lighthouse.2022_06_01_*`
+)
+
+
+SELECT
+  percentile,
+  client,
+  APPROX_QUANTILES(num_unsized_images, 1000)[OFFSET(percentile * 10)] AS num_unsized_images,
+  COUNTIF(num_unsized_images > 0) AS pages,
+  COUNT(0) AS total,
+  COUNTIF(num_unsized_images > 0) / COUNT(0) AS pct
+FROM
+  lh,
+  UNNEST([10, 25, 50, 75, 90, 100]) AS percentile
+GROUP BY
+  percentile,
+  client
+ORDER BY
+  percentile,
+  client

sql/2022/performance/fid_long_tasks.sql

@@ -0,0 +1,35 @@
+WITH long_tasks AS (
+  SELECT
+    _TABLE_SUFFIX AS client,
+    url AS page,
+    CAST(JSON_QUERY(item, '$.duration') AS FLOAT64) AS long_task_duration
+  FROM
+    `lighthouse.2022_06_01_*`,
+    UNNEST(JSON_QUERY_ARRAY(report, '$.audits.long-tasks.details.items')) AS item
+),
+
+per_page AS (
+  SELECT
+    client,
+    page,
+    SUM(long_task_duration) AS long_tasks
+  FROM
+    long_tasks
+  GROUP BY
+    client,
+    page
+)
+
+SELECT
+  percentile,
+  client,
+  APPROX_QUANTILES(long_tasks, 1000)[OFFSET(percentile * 10)] AS long_tasks
+FROM
+  per_page,
+  UNNEST([10, 25, 50, 75, 90, 100]) AS percentile
+GROUP BY
+  percentile,
+  client
+ORDER BY
+  percentile,
+  client

sql/2022/performance/fid_tbt.sql

@@ -0,0 +1,13 @@
+SELECT
+  percentile,
+  _TABLE_SUFFIX AS client,
+  APPROX_QUANTILES(CAST(JSON_QUERY(report, '$.audits.total-blocking-time.numericValue') AS FLOAT64), 1000)[OFFSET(percentile * 10)] AS tbt
+FROM
+  `httparchive.lighthouse.2022_06_01_*`,
+  UNNEST([10, 25, 50, 75, 90, 100]) AS percentile
+GROUP BY
+  percentile,
+  client
+ORDER BY
+  percentile,
+  client

sql/2022/performance/js_bytes_rank.sql

@@ -0,0 +1,15 @@
+SELECT
+  IF(_rank < 10000000, CAST(_rank AS STRING), 'all') AS rank,
+  _TABLE_SUFFIX AS client,
+  APPROX_QUANTILES(bytesJS, 1001)[OFFSET(501)] / 1024 AS js_kbytes
+FROM
+  `httparchive.summary_pages.2022_06_01_*`,
+  UNNEST([1000, 10000, 100000, 1000000, 10000000]) AS _rank
+WHERE
+  rank <= _rank
+GROUP BY
+  rank,
+  client
+ORDER BY
+  rank,
+  client

sql/2022/performance/lcp_bytes_distribution.sql

@@ -0,0 +1,36 @@
+WITH pages AS (
+  SELECT
+    _TABLE_SUFFIX AS client,
+    CAST(JSON_VALUE(payload, '$._metadata.page_id') AS INT64) AS pageid,
+    JSON_VALUE(payload, '$._performance.lcp_elem_stats.url') AS url
+  FROM
+    `httparchive.pages.2022_06_01_*`
+),
+
+requests AS (
+  SELECT
+    _TABLE_SUFFIX AS client,
+    pageid,
+    url,
+    respSize / 1024 AS kbytes
+  FROM
+    `httparchive.summary_requests.2022_06_01_*`
+)
+
+SELECT
+  percentile,
+  client,
+  APPROX_QUANTILES(kbytes, 1000)[OFFSET(percentile * 10)] AS kbytes
+FROM
+  pages
+JOIN
+  requests
+USING
+  (client, pageid, url),
+  UNNEST([10, 25, 50, 75, 90, 100]) AS percentile
+GROUP BY
+  percentile,
+  client
+ORDER BY
+  percentile,
+  client

sql/2022/performance/lcp_bytes_histogram.sql

@@ -0,0 +1,37 @@
+WITH pages AS (
+  SELECT
+    _TABLE_SUFFIX AS client,
+    CAST(JSON_VALUE(payload, '$._metadata.page_id') AS INT64) AS pageid,
+    JSON_VALUE(payload, '$._performance.lcp_elem_stats.url') AS url
+  FROM
+    `httparchive.pages.2022_06_01_*`
+),
+
+requests AS (
+  SELECT
+    _TABLE_SUFFIX AS client,
+    pageid,
+    url,
+    respSize / 1024 AS kbytes
+  FROM
+    `httparchive.summary_requests.2022_06_01_*`
+)
+
+SELECT
+  client,
+  IF(CEILING(kbytes / 100) * 100 < 1000, CAST(CEILING(kbytes / 100) * 100 AS STRING), '1000+') AS kbytes,
+  COUNT(0) AS freq,
+  SUM(COUNT(0)) OVER (PARTITION BY client) AS total,
+  COUNT(0) / SUM(COUNT(0)) OVER (PARTITION BY client) AS pct
+FROM
+  pages
+JOIN
+  requests
+USING
+  (client, pageid, url)
+GROUP BY
+  client,
+  kbytes
+ORDER BY
+  client,
+  kbytes

sql/2022/performance/lcp_format.sql

@@ -0,0 +1,36 @@
+WITH pages AS (
+  SELECT
+    _TABLE_SUFFIX AS client,
+    CAST(JSON_VALUE(payload, '$._metadata.page_id') AS INT64) AS pageid,
+    JSON_VALUE(payload, '$._performance.lcp_elem_stats.url') AS url
+  FROM
+    `httparchive.pages.2022_06_01_*`
+),
+
+requests AS (
+  SELECT
+    _TABLE_SUFFIX AS client,
+    pageid,
+    url,
+    format
+  FROM
+    `httparchive.summary_requests.2022_06_01_*`
+)
+
+SELECT
+  client,
+  format,
+  COUNT(0) AS freq,
+  SUM(COUNT(0)) OVER (PARTITION BY client) AS total,
+  COUNT(0) / SUM(COUNT(0)) OVER (PARTITION BY client) AS pct
+FROM
+  pages
+JOIN
+  requests
+USING
+  (client, pageid, url)
+GROUP BY
+  client,
+  format
+ORDER BY
+  pct DESC