[HUD] Keep going display on HUD: change queries to use temp conclusion + log result (#6845)

clee2000 · web-flow · commit 4e43bd8700fb · 2025-07-01T08:52:17.000-07:00
Change various queries to use some aliases that automatically handle keep going. Not every query needs to use this, only really things that need live data (ex commit page, HUD page, Dr. CI). There are probably some missing and this changes probably too many but I think it's ok for now I added an explanation of the table schema in #6848 Testing: `python tools/torchci/clickhouse_query_perf.py --base origin/main --results --strict-results --query hud_query` for each query Also added some tests for those that didn't have any. Changes: * log classifier count but it was because it takes different example lines that have slight differences (ex log lines include test time and test time changes) * recent_pr_workflows_query - I added a new logUrl result. There were some other differences but I think its just because timing I made a PR put keep-going on and purposely failed a job. While it continued to run, I checked that: * job shows up as failing * raw log link goes to temp log * log viewer shows temp log * dr ci showed it failed
diff --git a/torchci/clickhouse_queries/commit_failed_jobs/query.sql b/torchci/clickhouse_queries/commit_failed_jobs/query.sql
@@ -16,19 +16,19 @@ SELECT
   CONCAT(j.workflow_name, ' / ', j.name) AS name,
   j.runner_name AS runnerName,
   w.authorEmail as authorEmail,
-  j.conclusion as conclusion,
+  j.conclusion_kg as conclusion,
   j.completed_at as completed_at,
   j.html_url as html_url,
   j.head_sha as head_sha,
   w.head_sha_timestamp AS head_sha_timestamp,
   j.head_branch as head_branch,
-  j.torchci_classification.'captures' AS failure_captures,
-  IF(j.torchci_classification.'line' = '', [], [j.torchci_classification.'line']) AS failure_lines,
-  j.torchci_classification.'context' AS failure_context,
+  j.torchci_classification_kg.'captures' AS failure_captures,
+  IF(j.torchci_classification_kg.'line' = '', [], [j.torchci_classification_kg.'line']) AS failure_lines,
+  j.torchci_classification_kg.'context' AS failure_context,
   j.created_at AS time
 FROM
   default.workflow_job j final
   JOIN runs w on w.id = j.run_id
 WHERE
   j.id in (select id from materialized_views.workflow_job_by_head_sha where head_sha in {shas: Array(String)})
-  AND j.conclusion IN ('failure', 'cancelled')
+  AND j.conclusion_kg IN ('failure', 'cancelled')
diff --git a/torchci/clickhouse_queries/commit_jobs_batch_query/params.json b/torchci/clickhouse_queries/commit_jobs_batch_query/params.json
@@ -2,5 +2,30 @@
   "params": {
     "shas": "Array(String)"
   },
-  "tests": []
-}
+  "tests": [
+    {
+      "shas": [
+        "c808af514d59e25ea4a880c1e1e07d3232984e5d",
+        "2f94f69b7c83370ef0cc65e3ab96bb5bf11a7b1a",
+        "197c1869f5467b4cfa1e197bbe394be82ba40554",
+        "2d39a48d524021995269411bd49fe792e59d9f94",
+        "53e0b9c3936176521ed8d71c00abd5b7499057c2",
+        "de45c5f673ce261e9a82c54280beeda36cff640e",
+        "18b01afa9ed1dbc696a06de1b69bf5c021c18c10",
+        "bbf1a6feac0e88772be103a1b159b871b5c00b4a",
+        "455dfd258980294f0745bd90aee12a323e37224d",
+        "50b2069b61942e923528c94ccbbc8ab5e92c381e",
+        "dfc31b3345d78b0a49d446dcc1957404606a3aa2",
+        "0d01bafc34fc99a0b3e46cbf1ecfd8f97563efa6",
+        "127695eb5c973f9fdba24c47b465e30a19292582",
+        "0a16818d5b3fdf0fb8148dd6b849687251a56376",
+        "98e594b565471c2dbf9f50fe624259741efbad61",
+        "a730c65fe3d4dd6141f69720a70440cd4debc6d2",
+        "4585c33e74079af8e9067bf39970b93c3f13629f",
+        "7521cd91118c1c1a55b8f7146d7e22bf60fe06bf",
+        "68e023cbbb39b46ecc607329bb062db5c48ac196",
+        "df9e5a276b47086aaddb04885606bdcdbf87b593"
+      ]
+    }
+  ]
+}
diff --git a/torchci/clickhouse_queries/commit_jobs_batch_query/query.sql b/torchci/clickhouse_queries/commit_jobs_batch_query/query.sql
@@ -10,21 +10,18 @@ WITH job AS (
         job.id as id,
         workflow.id AS workflow_id,
         workflow.artifacts_url AS github_artifact_url,
-        job.conclusion as conclusion,
+        job.conclusion_kg as conclusion,
         job.html_url as html_url,
-        CONCAT(
-            'https://ossci-raw-job-status.s3.amazonaws.com/log/',
-            job.id
-        ) AS log_url,
+        job.log_url as log_url,
         DATE_DIFF(
             'SECOND',
             job.started_at,
             job.completed_at
         ) AS duration_s,
-        IF(job.torchci_classification.'line' = '', [], [job.torchci_classification.'line']) AS failure_line,
-        job.torchci_classification.'context' AS failure_context,
-        job.torchci_classification.'captures' AS failure_captures,
-        job.torchci_classification.'line_num' AS failure_line_number
+        IF(job.torchci_classification_kg.'line' = '', [], [job.torchci_classification_kg.'line']) AS failure_line,
+        job.torchci_classification_kg.'context' AS failure_context,
+        job.torchci_classification_kg.'captures' AS failure_captures,
+        job.torchci_classification_kg.'line_num' AS failure_line_number
     FROM
         default.workflow_job job final
         INNER JOIN default.workflow_run workflow final ON workflow.id = job.run_id
diff --git a/torchci/clickhouse_queries/commit_jobs_query/params.json b/torchci/clickhouse_queries/commit_jobs_query/params.json
@@ -3,5 +3,10 @@
     "repo": "String",
     "sha": "String"
   },
-  "tests": []
-}
+  "tests": [
+    {
+      "repo": "pytorch/pytorch",
+      "sha": "85df746892d9b0e87e7a5dfa78ef81a84aec6de0"
+    }
+  ]
+}
diff --git a/torchci/clickhouse_queries/commit_jobs_query/query.sql b/torchci/clickhouse_queries/commit_jobs_query/query.sql
@@ -12,27 +12,15 @@ WITH job AS (
         workflow.id AS workflow_id,
         workflow.artifacts_url AS github_artifact_url,
         multiIf(
-            job.conclusion = ''
+            job.conclusion_kg = ''
             and status = 'queued' ,
             'queued',
-            job.conclusion = '',
+            job.conclusion_kg = '',
             'pending',
-            job.conclusion
+            job.conclusion_kg
         ) as conclusion,
         job.html_url,
-        IF(
-            {repo: String } = 'pytorch/pytorch',
-            CONCAT(
-                'https://ossci-raw-job-status.s3.amazonaws.com/log/',
-                job.id:: String
-            ),
-            CONCAT(
-                'https://ossci-raw-job-status.s3.amazonaws.com/log/',
-                {repo: String },
-                '/',
-                job.id:: String
-            )
-        ) AS log_url,
+        job.log_url AS log_url,
         if(
             job.started_at = 0,
             0,
@@ -43,10 +31,10 @@ WITH job AS (
             0,
             DATE_DIFF('SECOND', job.started_at, job.completed_at)
         ) AS duration_s,
-        job.torchci_classification.line as line,
-        job.torchci_classification.captures as captures,
-        job.torchci_classification.line_num as line_num,
-        job.torchci_classification.context as context,
+        job.torchci_classification_kg.'line' as line,
+        job.torchci_classification_kg.'captures' as captures,
+        job.torchci_classification_kg.'line_num' as line_num,
+        job.torchci_classification_kg.'context' as context,
         job.runner_name AS runner_name,
         workflow.head_commit. 'author'.'email' AS authorEmail
     FROM
diff --git a/torchci/clickhouse_queries/failed_workflow_jobs/params.json b/torchci/clickhouse_queries/failed_workflow_jobs/params.json
@@ -5,5 +5,12 @@
     "startTime": "DateTime64(3)",
     "stopTime": "DateTime64(3)"
   },
-  "tests": []
-}
+  "tests": [
+    {
+      "branch": "main",
+      "repo": "pytorch/pytorch",
+      "startTime": "2025-01-01T00:00:00.000",
+      "stopTime": "2025-01-15T00:00:00.000"
+    }
+  ]
+}
diff --git a/torchci/clickhouse_queries/failed_workflow_jobs/query.sql b/torchci/clickhouse_queries/failed_workflow_jobs/query.sql
@@ -3,16 +3,13 @@ SELECT DISTINCT
     j.head_sha AS sha,
     CONCAT(w.name, ' / ', j.name) AS jobName,
     j.id,
-    j.conclusion,
+    j.conclusion_kg as conclusion,
     j.html_url AS htmlUrl,
-    CONCAT(
-        'https://ossci-raw-job-status.s3.amazonaws.com/log/',
-        j.id
-    ) AS logUrl,
+    j.log_url as logUrl,
     DATE_DIFF('SECOND', j.started_at, j.completed_at) AS durationS,
-    array(j.torchci_classification. 'line') AS failureLines,
-    j.torchci_classification. 'captures' AS failureCaptures,
-    array(j.torchci_classification. 'line_num') AS failureLineNumbers
+    array(j.torchci_classification_kg. 'line') AS failureLines,
+    j.torchci_classification_kg. 'captures' AS failureCaptures,
+    array(j.torchci_classification_kg. 'line_num') AS failureLineNumbers
 FROM
     workflow_job j FINAL
     JOIN workflow_run w FINAL on w.id = j.run_id
@@ -23,4 +20,4 @@ WHERE
     AND w.head_branch = {branch: String }
     AND w.event != 'workflow_run'
     AND w.event != 'repository_dispatch'
-    AND j.conclusion IN ('failure', 'cancelled', 'time_out')
+    AND j.conclusion_kg IN ('failure', 'cancelled', 'time_out')
diff --git a/torchci/clickhouse_queries/flaky_tests/ind_info/query.sql b/torchci/clickhouse_queries/flaky_tests/ind_info/query.sql
@@ -16,15 +16,15 @@ WITH failed_test_runs AS (
 
 failed_jobs AS (
     SELECT
-        j.conclusion AS conclusion,
+        j.conclusion_kg AS conclusion,
         j.id AS id,
         j.run_id AS run_id,
         j.name AS name,
         j.html_url AS html_url,
         j.started_at AS started_at,
-        tupleElement(j.torchci_classification, 'line') AS line,
-        tupleElement(j.torchci_classification, 'line_num') AS line_num,
-        tupleElement(j.torchci_classification, 'captures') AS captures,
+        tupleElement(j.torchci_classification_kg, 'line') AS line,
+        tupleElement(j.torchci_classification_kg, 'line_num') AS line_num,
+        tupleElement(j.torchci_classification_kg, 'captures') AS captures,
         j.head_sha AS head_sha
     FROM default.workflow_job AS j
     WHERE
diff --git a/torchci/clickhouse_queries/hud_query/query.sql b/torchci/clickhouse_queries/hud_query/query.sql
@@ -5,30 +5,18 @@ WITH job AS (
         job.workflow_name as workflow_name,
         job.id as id,
         job.status as status,
-        job.conclusion as conclusion,
+        job.conclusion_kg as conclusion,
         job.html_url as html_url,
-        IF(
-          {repo: String} = 'pytorch/pytorch',
-          CONCAT(
-              'https://ossci-raw-job-status.s3.amazonaws.com/log/',
-              job.id::String
-            ),
-          CONCAT(
-              'https://ossci-raw-job-status.s3.amazonaws.com/log/',
-              {repo: String},
-              '/',
-              job.id::String
-            )
-          ) as log_url,
+        job.log_url as log_url,
         if(
             job.completed_at = 0,
             null,
             DATE_DIFF('SECOND', job.started_at, job.completed_at)
         ) AS duration_s,
         job.repository_full_name as repo,
-        job.torchci_classification.'line' as line,
-        job.torchci_classification.'captures' as captures,
-        job.torchci_classification.'line_num' as line_num,
+        job.torchci_classification_kg.'line' as line,
+        job.torchci_classification_kg.'captures' as captures,
+        job.torchci_classification_kg.'line_num' as line_num,
         annotation.annotation as annotation
     FROM
         workflow_job job final
diff --git a/torchci/clickhouse_queries/log_captures_count/params.json b/torchci/clickhouse_queries/log_captures_count/params.json
@@ -3,5 +3,10 @@
     "startTime": "DateTime64(3)",
     "stopTime": "DateTime64(3)"
   },
-  "tests": []
-}
+  "tests": [
+    {
+      "startTime": "2025-01-01T00:00:00.000",
+      "stopTime": "2025-01-15T00:00:00.000"
+    }
+  ]
+}
diff --git a/torchci/clickhouse_queries/log_captures_count/query.sql b/torchci/clickhouse_queries/log_captures_count/query.sql
@@ -1,7 +1,7 @@
 WITH jobs AS (
     SELECT
-        j.torchci_classification.line AS line,
-        j.torchci_classification.captures AS captures,
+        j.torchci_classification_kg.'line' AS line,
+        j.torchci_classification_kg.'captures' AS captures,
         j.run_id
     FROM
         default.workflow_job j FINAL
@@ -12,7 +12,7 @@ WITH jobs AS (
                 created_at >= {startTime: DateTime64(3)}
                 AND created_at < {stopTime: DateTime64(3)}
         )
-        AND j.conclusion IN ('cancelled', 'failure', 'time_out')
+        AND j.conclusion_kg IN ('cancelled', 'failure', 'time_out')
 )
 
 SELECT
diff --git a/torchci/clickhouse_queries/recent_pr_workflows_query/query.sql b/torchci/clickhouse_queries/recent_pr_workflows_query/query.sql
@@ -44,20 +44,21 @@ SELECT
   w.head_commit.'author'.'email' as authorEmail,
   CONCAT(w.name, ' / ', j.name) AS name,
   j.name AS jobName,
-  j.conclusion as conclusion,
+  j.conclusion_kg as conclusion,
   j.completed_at as completed_at,
   j.html_url as html_url,
+  j.log_url as logUrl,
   j.head_branch as head_branch,
   recent_prs.number AS pr_number,
   recent_prs.sha AS head_sha,
   recent_prs.timestamp AS head_sha_timestamp,
-  j.torchci_classification.'captures' AS failure_captures,
+  j.torchci_classification_kg.'captures' AS failure_captures,
   IF(
-    j.torchci_classification.'line' = '',
+    j.torchci_classification_kg.'line' = '',
     [],
-    [j.torchci_classification.'line']
+    [j.torchci_classification_kg.'line']
   ) AS failure_lines,
-  j.torchci_classification.'context' AS failure_context,
+  j.torchci_classification_kg.'context' AS failure_context,
   j.created_at AS time
 FROM
   default.workflow_run w final
@@ -78,6 +79,7 @@ SELECT
   w.conclusion as conclusion,
   toDateTime64(0, 9) as completed_at,
   w.html_url as html_url,
+  '' as logUrl,
   w.head_branch as head_branch,
   recent_prs.number AS pr_number,
   w.head_sha AS head_sha,
diff --git a/torchci/lib/jobUtils.ts b/torchci/lib/jobUtils.ts
@@ -235,6 +235,10 @@ export function removeJobNameSuffix(
 }
 
 export async function hasS3Log(job: RecentWorkflowsData): Promise<boolean> {
+  if (job.logUrl !== undefined && job.logUrl !== "") {
+    const res = await fetch(job.logUrl, { method: "HEAD" });
+    return res.status !== 404;
+  }
   // This is to handle the infra flaky issue where the log is not available on
   // S3 and no failure is found.
   // NB: PyTorch uses the shortcut /log/JOB_ID path while other repos require
diff --git a/torchci/lib/types.ts b/torchci/lib/types.ts
@@ -46,6 +46,7 @@ export interface RecentWorkflowsData extends BasicJobData {
   id: number;
   completed_at: string;
   html_url: string;
+  logUrl?: string;
   head_sha: string;
   head_sha_timestamp: string;
   head_branch: string;
