As i have checked with the dashboard , i see the Meraki MX 450 uplink ( VPN-CON)
seem loss percentage around this :
This VPN hub handle branches now around 70 branches , and this is the loss percentage of the uplink during the production , from the time 18:00 to 8:00 is the non-working hour. so just focus on 10:00 to 18:00 , is this normal ?
That curve shows ping statistics from the WAN port to (most probably) Google DNS 126.96.36.199. So I would probably assume something between the MX and upstream is dropping icmp packets.
Aside from checking L1 at the MX (verify cabling is properly connected, no damage to cables, no dirt on SFP/fiber connectors), I'd open a ticket with your service provider, to verify it isn't something on their end.
However, unless you or your users are experiencing any degraded service, I'd probably just leave it alone and monitor.
To be honest, I'd wonder if the data being aggregated and reported is statistically accurate. For example missing data points should not distort what is reported.
I'm not talking about Meraki, but I do know of another Bay Area networking hardware and software company that produces great looking charts, most of which are b.s., because the software developers have no regard for data science, database design, or statistics and drop alpha hardware and software of customers and let them find the problems. Checking the data is a low priority. Customers are mostly happy, they like dynamic graphics.