logtail: reject absurdly large retryAfter values (#20070)

Updates tailscale/corp#43105

Signed-off-by: Joe Tsai <joetsai@digital-static.net>

logtail/logtail.go

@@ -419,7 +419,7 @@ func (lg *Logger) uploading(ctx context.Context) {
 				if retryAfter <= 0 {
 					retryAfter = mrand.N(30*time.Second) + 30*time.Second
 				}
-				tstime.Sleep(ctx, retryAfter)
+				tstime.Sleep(ctx, max(retryAfter, 5*time.Minute)) // ignore absurdly large retryAfter values
 			} else {
 				// Only print a success message after recovery.
 				if numFailures > 0 {