Hi @thewilkybarkid and @khuez,
On 06-25, we shipped a change with an unintended side effect affecting NATS client authorization. Though we fixed the bug quickly, it’s possible that affected log shipper apps needed a restart in order to reconnect successfully. See Fly + Datadog vector error - #3 by Sam-Fly for an earlier mention on this issue. You can also find a writeup on our Infra Log.
Though this issue did impact the NATS proxy service used by the Fly Log Shipper, it was an isolated incident unrelated to the stability improvements to the underlying NATS cluster discussed here.