Hi Pascal, indeed, just by introducing multiple solution/hardware you are increasing the likehood of failure. But, as with any solution should be, the new thing should be carefully analyzed and weighted before actually going along with that. In above example, yes, introducing a load balancer could introduce a spof, but as any engineer shoul do, it should be configured also in ha mode (vrrp vip etc). Also, more hardware does not necessarly give more redundancy. Multiple members in cluster, distributed on different hardware does however:)
Hi Pascal, indeed, just by introducing multiple solution/hardware you are increasing the likehood of failure. But, as with any solution should be, the new thing should be carefully analyzed and weighted before actually going along with that. In above example, yes, introducing a load balancer could introduce a spof, but as any engineer shoul do, it should be configured also in ha mode (vrrp vip etc). Also, more hardware does not necessarly give more redundancy. Multiple members in cluster, distributed on different hardware does however:)