慢查询-GAE PHP和Cloud SQL

Question

I have a scheduling app that allows users to create events in bulk using a weekly template. This works by posting the template lessons via AJAX to a PHP file that takes the template and replicates it the desired number of times. The number of events it has to create at one time could be from 200-800 depending.

The problem I'm having is that these queries take a long time to finish, which causes the website to hang and sometimes time out while waiting for all the entries to be made.

A simplified example:

$events = array('valid array of all events and their details');
foreach($events as $event) {
      $query = 'Valid SQL with individual event detaisl';
      mysqli_query($connection, $query);
}

In reality more would be going on, for example, calculating start and end times, associating users from junction tables , etc. This would have to run 200-800 times to create an entry for each individual event made by the template.

How would I go about speeding this up?

Any insight is greatly appreciated!

UPDATE: I've already tried increasing the Tier of the SQL instance. It doesn't have any effect on the amount of time it takes.

Answer 1

Some suggestions to consider:

1. Creating 200-800 new database entries per single request sounds a bit excessive. The logic of your app and the data models/structure you use could probably be significantly improved and optimized by not creating all events at once, instead, only saving the settings that let you generate the 800 entries with some code that would return an individual event if needed and save it separately if any custom (not from the template) info was added or changed in the future. This would probably require a lot of changes in your code and might or might not be worth it (time-wise not speed-wise which would definitely improve).

2. As already mentioned in the comments, this could be taken into a separate task with some UI feedback to the user saying that the events are being created, and then another notification saying that it is now done.

3. By looking at your code it looks like (I am not very familiar with PHP/mysql) when creating 800 events you also contact the database 800 separate times (each time waiting for a completion response). Assuming that the bottleneck was in these multiple connects then by concatenating your generated queries and feeding them to the database all at once you could potentially achieve some performance improvements. See example below (note the required semicolon at the end of each generated query and the mysqli_multi_query function allowing sending multiple queries at once):

   $events = array('valid array of all events and their details'); $query = ""; foreach($events as $event) { $query .= 'Valid SQL with individual event detaisl;\\n'; } mysqli_multi_query($connection, $query);