Hello @seljicom !
Yes you were correct, with the last plugin updates I was trying to solve a problem related with the automatic updates of product data, however on some website the process get stuck and I don’t have found solution yet.
Did you have CRON related problems before the Creators API switch?
Did you already try to mass update?
Are you using the Creators API v2 or v3?
Thread Starter
JSN
(@seljicom)
Got it. So I’m pretty sure my CRON issues had started a while ago, likely prior to the switch to the new API. I recall noticing “N number of scheduled events are overdue” all the time inside my Updraft backup plugin at some point. On the API version, just checked and I was on the (old) 2.1, so will try switching to the latest then. I did try to run the mass update initially but I think it updated some small number and then stopped (screenshot reflects the result).
Ok, so it seems that your CRON related problems are not directly related with the new APIs.
The mass update is an experimental feature i’m developing to mass update the database when the normal cron updates fails. It is still raw. With certain problematic products it can stop abruptly. Recently I’ve limited the mass update size to just 100 products to prevent crashes, because it actually crashes… It should be used only to update huge product databases in short times .
Now you can:
1) Try to give a try with the 3.1 credentials
2) Try to see if the single product update works (i.e. clicking the green circling arrow)
3) Temporary turn off the Amazing Affiliates automatic updates in the setting and check if you still got CRON related errors, if yes, we might have found the culprit in the plugin CRON script :/
Thread Starter
JSN
(@seljicom)
A few things I’ve discovered:
1. Once switched over to v3.1 I forgot to create a new set of credentials on Amazon (ideally switching API versions should be invalidating the API health check)
2. Although my Auto Update was on, it was showing as off in the debug screen, which I wasn’t paying too much attention to. So I went ahead and flipped on/of while saving and now it shows every 2 minutes.
One thing I don’t understand is how the scheduler meant to work. Let’s say I’ve set to auto update every 2 mins and I have 500 products that need updates. Would the auto feature send batches of ASINs to be updated every 2 mins, and make sure all 500 eventually get updated i.e. find the next stale batch and continue cycling 1-500 infinitely?
Will keep observing. (btw individual updates work just fine, the green circular arrow)
-
This reply was modified 2 months, 2 weeks ago by
JSN.
Well at least the individual updates still works, thank you for the feedback.
As for:
1. Once switched over to v3.1 I forgot to create a new set of credentials on Amazon (ideally switching API versions should be invalidating the API health check)
Good suggestion, taking notes, thank you.
2. Although my Auto Update was on, it was showing as off in the debug screen, which I wasn’t paying too much attention to. So I went ahead and flipped on/of while saving and now it shows every 2 minutes.
As a general life rule, it is always a good idea to switch settings when things get stuck.
Does it work now? Does it try to update more frequently?
One thing I don’t understand is how the scheduler meant to work. Let’s say I’ve set to auto update every 2 mins and I have 500 products that need updates. Would the auto feature send batches of ASINs to be updated every 2 mins, and make sure all 500 eventually get updated i.e. find the next stale batch and continue cycling 1-500 infinitely?
When it works, it is simple.
A CRON event periodically (in your case every 120 seconds) picks the 10 least updated products from your database and makes an API call to get the updated info of those 10.
Why 10? It is due to how the Amazon APIs are designed, they can take up to 10 ASIN per single call, hence every auto update cycle consume just a single call (of the 8640 available per day).
At the moment it is designed to be the most efficient, but in future I may change or fork it in two versions as in some website with big databases this process gets stuck.
