AWS FusionHub VPN flapping every couple seconds

This is my topology

I have PepVPN running on OSPF Area 0.0.0.0 and mutually redistributing BGP routes I learn from Transit Gateway (AWS) via Transit Gateway Connect and also impoting AWS routes.

When it create the profile it works well but after 200 seconds (sometimes 400s ) it peplink onpremises detects a link failure and sometimes it is able to reestablish after several attempts but it seams to happen randomly.

I have all three appliances in Ip Forwarding mode but that that doesn’t seem to make a difference from what I can tell, any idea what might going on?

This is important to me as I want to build a fully redundant scenario on AWS (redundancy doesn’t seem to be a problem) as when everything is up, no routing loops or whatsoever.

Only thing the logs tell me is :

Jul 25 20:02:03 SpeedFusion: conn_to_AWS-Pepe-1 failed to establish connection
Jul 25 20:01:47 SpeedFusion: conn_to_AWS-Pepe-1: Initiated TLSv1.3 connection to 34.228.150.112 using cipher suite TLS_AES_256_GCM_SHA384
Jul 25 20:01:46 SpeedFusion: conn_to_AWS-Pepe-2 failed to establish connection (failed to connect remote peer)
Jul 25 20:00:56 SpeedFusion: conn_to_AWS-Pepe-1 failed to establish connection
Jul 25 20:00:15 SpeedFusion: conn_to_AWS-Pepe-1 failed to establish connection (failed to connect remote peer)
Jul 25 19:55:55 SpeedFusion: conn_to_AWS-Pepe-1 failed to establish connection
Jul 25 19:55:39 SpeedFusion: conn_to_AWS-Pepe-1: Initiated TLSv1.3 connection to 34.x,x.x using cipher suite TLS_AES_256_GCM_SHA384
Jul 25 19:55:29 SpeedFusion: AWS-A (AWS-Pepe-1, sn:1191-C211-603E) disconnected from conn_to_AWS-Pepe-1 (link failure detected)
Jul 25 19:55:15 SpeedFusion: conn_to_AWS-Pepe-2 failed to establish connection (failed to connect remote peer)
Jul 25 19:55:04 SpeedFusion: AWS-Pepe2 (AWS-Pepe-2, sn:111D-51BB-38A7) disconnected from conn_to_AWS-Pepe-2 (link failure detected)
Jul 25 19:46:37 SpeedFusion: AWS-A (AWS-Pepe-1, sn:1191-C211-603E) connected to conn_to_AWS-Pepe-1
Jul 25 19:45:15 SpeedFusion: AWS-Pepe2 (AWS-Pepe-2, sn:111D-51BB-38A7) connected to conn_to_AWS-Pepe-2
Jul 25 19:44:41 SpeedFusion: conn_to_AWS-Pepe-2 failed to establish connection
Jul 25 19:44:39 SpeedFusion: conn_to_AWS-Pepe-2: Initiated TLSv1.3 connection to 18.x.x.x using cipher suite TLS_AES_256_GCM_SHA384
Jul 25 19:44:23 SpeedFusion: conn_to_AWS-Pepe-1 failed to establish connection
Jul 25 19:44:21 SpeedFusion: conn_to_AWS-Pepe-1: Initiated TLSv1.3 connection to 34.x.x.x using cipher suite TLS_AES_256_GCM_SHA384
Jul 25 19:43:49 SpeedFusion: AWS-A (AWS-Pepe-1, sn:1191-C211-603E) disconnected from conn_to_AWS-Pepe-1 (link failure detected)
Jul 25 19:42:55 SpeedFusion: conn_to_AWS-Pepe-2 failed to establish connection (failed to connect remote peer)
Jul 25 19:42:44 SpeedFusion: AWS-Pepe2 (AWS-Pepe-2, sn:111D-51BB-38A7) disconnected from conn_to_AWS-Pepe-2 (link failure detected)
Jul 25 19:33:52 SpeedFusion: AWS-A (AWS-Pepe-1, sn:1191-C211-603E) connected to conn_to_AWS-Pepe-1
Jul 25 19:33:39 SpeedFusion: AWS-Pepe2 (AWS-Pepe-2, sn:111D-51BB-38A7) connected to conn_to_AWS-Pepe-2
Jul 25 19:33:21 SpeedFusion: conn_to_AWS-Pepe-1 failed to establish connection
Jul 25 19:33:20 SpeedFusion: conn_to_AWS-Pepe-2: Initiated TLSv1.3 connection to 18.x.x.x using cipher suite TLS_AES_256_GCM_SHA384
Jul 25 19:33:19 SpeedFusion: conn_to_AWS-Pepe-1: Initiated TLSv1.3 connection to 34.x.x.x using cipher suite TLS_AES_256_GCM_SHA384

Everytime I rebuild the profiles on Inconnect2 it immediately comes online.

I’ve enabled an option and it seems stable now.

I have not enabled it before since I was manually redistributing in BGP & OSPF Settings.