Pivot tables returning odd values after running scheduled analytics

Users have been complaining that after running the analytics (we run it at specific times everyday) the pivot tables show odd values when looking at past entered data in tracker.
We have implemented a tracker that users enter data daily. We also have created program indicators that allow the users to pivot tables to do reports.
We run analytics 3 times a day (01AM, 11AM and 02PM), this is scheduled in Scheduler app. In the server (Ubuntu Server 18.04), we created a cronjob that clears the analytics table 5 minutes before Dhis2 runs the scheduled analytics

The issue is, after running the analytics, the users complain that the data for previous dates are odd:
For instance let us suppose we are on a Friday, if we check the program indicators for that data inserted on Monday, and compare it to what we saw on Tuesday, Wednesday and so on, but for Monday data, the returned values are very different.
But this is solved if we mannualy clear the analytics table and run analytics.
It seems to be an issue with the scheduled analytics.
I went to the terminal to check if everything is working accordingly and as far as I can say, the logs for analytics do not return an error.
I’m kinda lost on what to do because it seems everythings is functioning perfectly but the users keep complaining.

I’m using DHIS2 2.33.9

Best Regards

Dear @joao.machiana welcome to the community!

Does the superuser face the same issue? For the users that complain about odd values, would it be possible to ask them to clear the browser cache first? It might be a cache issue.

What happens if you manually run the analytics tables without first ‘clearing the analytics tables’?

This got fixed. I cannot tell you the sollution because I don’t know also (laughs).
One thing that I did was to drastically increase server space (previously had around 5GB free space), to have something like 20~30GB free space.

We got the running system from another company, and they were running the analytics like that (clearing the analytics first and then running it). They said that when developing the system they were having some issues with program indicators, and were recomended that to avoid issues it was better to perform a cleanup on the analytics and the running it.

Thanks for help