SharePoint isn't available at our office network for "random" time frames

[YvonneP]

We have this issue since nearly 2 years now and changed a lot in our infrastructure, got a new firewall, new switches and better bandwith, but it stayed the same.
We noticed that those errors don't happen during the weekend, when no one is in the office and became less frequent during corona, probably because many workers are now working from home.
Sites like www.office.com are still available fine during those fallouts, only our sharepoint page is not. The fallouts can be only a few seconds or even half an hour.

Information that is probably relevant:

• We use SharePoint online, not on-prem
• OneDrive also doesn't work during the fallouts
• Other Office products like MS Teams do work
• The SharePoint page is still available outside from our network
• I can still ping and use trcrt with the IP during the fallouts, but can't reach it with telnet
• Changing the DNS from the router didn't work
• During the fallouts, we noticed long TCP Keep-Alive and Retransmission sequences while monitoring with wireshark

What would your approach be for this problem?

[#2]

@YvonneP 

• During the fallouts, we noticed long TCP Keep-Alive and Retransmission sequences while monitoring with wireshark

Is that for all sites/traffic or just SharePoint?  What does network utilisation look like at this time? 

Check the Microsoft 365 usage reports in the admin centre and check if anyone is syncing large amounts of data/number of files for SharePoint or OneDrive. Do you tell you users to sync document libraries as a default or only the files they need? 

I had a query the other day where the customer had 1 document library, 300GB and 320,000 files and all user were syncing the data and it was killing every laptop. They were doing this because they always used to have every file accessible in file explorer.