Hi everyone, I've been running a proof-of-concept with Versa Networks for about three months now. Overall, the setup was smooth, but I'm hitting a specific issue I can't seem to debug.
We have a site-to-site tunnel between our main AWS VPC and a colocation facility. Whenever we push large batch data transfers (think multi-gigabyte database backups), the tunnel will intermittently drop. The connection re-establishes automatically after 30-60 seconds, but it's enough to kill the transfer job. The puzzling part is that the Versa Director logs don't show any obvious errors—no policy violations, no resource alerts, just a clean disconnect and reconnect.
Here's what I've checked so far:
* Both ends have identical, conservative IPSec parameters (SHA2-256, AES-256, DH group 14).
* The underlay connections (ISP links) show no packet loss or saturation according to the Versa analytics.
* CPU and memory utilization on the Versa appliances (we're using the virtual FlexVNF) spike during the transfer but stay within "normal" bounds per their docs.
My main questions are:
1. Has anyone else experienced silent drops under heavy load? Is this a known throttling or session limit behavior?
2. How does Versa's tunnel stability compare under load to other vendors like Cisco Viptela or Fortinet? I've heard some platforms handle rekeying during high throughput better.
3. Are there specific, less-obvious logs (maybe kernel or forwarding plane) I should be looking at on the Versa device itself, rather than in Director?
I'm trying to figure out if this is a misconfiguration on our end, a platform limitation, or something we need to take to Versa support. Any insights from longer-term users would be really helpful.