Group jobs by hour in SQL rather than PHP

This makes a big reduction in the number of rows returned
by our SQL queries.

app/Http/Controllers/MonitorController.php

@@ -10,6 +10,7 @@
 use App\Enums\ClassicPalette;
 use App\Enums\HighContrastPalette;
 use Illuminate\Http\JsonResponse;
+use Illuminate\Support\Collection;
 use Illuminate\Support\Facades\Auth;
 use Illuminate\Support\Facades\DB;
 use Symfony\Component\HttpFoundation\Response;
@@ -50,10 +51,10 @@ public function get(): JsonResponse
  new DateTime('24 hours ago'),
  new DateInterval('PT1H'),
  24);
-
  $i = 0;
  foreach ($period as $datetime) {
- $timestamp = $this->truncateTimestampToHour($datetime);
+ // Truncate this timestamp to the beginning of the hour.
+ $timestamp = $datetime->setTime(intval($datetime->format('H')), 0, 0)->getTimestamp();
  $success_values[$timestamp] = 0;
  $fail_values[$timestamp] = 0;
  $i++;
@@ -63,20 +64,21 @@ public function get(): JsonResponse
  }
 
  // Populate pass/fail trendline data.
- foreach (DB::table('failed_jobs')->pluck('failed_at') as $failed_at) {
- $key = $this->truncateTimestampToHour(new DateTime($failed_at));
+ foreach ($this->resultsPerHour('failed_jobs', 'failed_at') as $row) {
+ $key = $row->truncated_time;
  if (array_key_exists($key, $fail_values)) {
- $fail_values[$key] += 1;
+ $fail_values[$key] += $row->n_jobs;
  } else {
- $fail_values[$key] = 1;
+ $fail_values[$key] = $row->n_jobs;
  }
  }
- foreach (DB::table('successful_jobs')->pluck('finished_at') as $finished_at) {
- $key = $this->truncateTimestampToHour(new DateTime($finished_at));
+
+ foreach ($this->resultsPerHour('successful_jobs', 'finished_at') as $row) {
+ $key = $row->truncated_time;
  if (array_key_exists($key, $success_values)) {
- $success_values[$key] += 1;
+ $success_values[$key] += $row->n_jobs;
  } else {
- $success_values[$key] = 1;
+ $success_values[$key] = $row->n_jobs;
  }
  }
 
@@ -110,10 +112,62 @@ public function get(): JsonResponse
  }
 
  /**
- * Return the timestamp of the beginning of the hour for a given timestamp.
+ * Group timestamp values by hour.
+ * @return Collection<int,\stdClass>
+ */
+ private function resultsPerHour(string $table, string $field) : Collection
+ {
+ if (config('database.default') === 'mysql') {
+ return $this->mySQLResultsPerHour($table, $field);
+ } else {
+ return $this->postgreSQLResultsPerHour($table, $field);
+ }
+ }
+
+ /**
+ * MySQL implementation of resultsPerHour
+ * @return Collection<int,\stdClass>
+ */
+ private function mySQLResultsPerHour(string $table, string $field) : Collection
+ {
+ // Group jobs by hour.
+ // We achieve this by:
+ // 1) subtracting the seconds
+ // 2) subracting the minutes
+ // 3) casting the result to a UNIX timestamp
+ return DB::table($table)
+ ->select(DB::raw("
+ UNIX_TIMESTAMP(
+ DATE_SUB(
+ DATE_SUB({$field}, INTERVAL MINUTE({$field}) MINUTE),
+ INTERVAL SECOND({$field}) SECOND
+ )
+ ) AS truncated_time,
+ COUNT(1) AS n_jobs
+ "))
+ ->groupBy('truncated_time')
+ ->get();
+ }
+
+ /**
+ * Postgres implementation of resultsPerHour
+ * @return Collection<int,\stdClass>
  */
- private function truncateTimestampToHour(DateTime $datetime) : int
+ private function postgreSQLResultsPerHour(string $table, string $field) : Collection
  {
- return $datetime->setTime(intval($datetime->format('H')), 0, 0)->getTimestamp();
+ // Group jobs by hour.
+ // We achieve this by:
+ // 1) using DATE_TRUNC() to truncate timestamps to the hour
+ // 2) using EXTRACT(EPOCH ...) to convert this value to a UNIX timestamp
+ $timezone = date_default_timezone_get();
+ return DB::table($table)
+ ->select(DB::raw("
+ EXTRACT(EPOCH FROM
+ DATE_TRUNC('hour', {$field}) AT TIME ZONE '{$timezone}'
+ )::INTEGER AS truncated_time,
+ COUNT(1) AS n_jobs
+ "))
+ ->groupBy('truncated_time')
+ ->get();
  }
 }