Reverse Proxy
A server that receives client requests on behalf of one or more backends, forwarding them and returning the response.
The single most useful placement in a web architecture, because it is one place to put many cross-cutting things: TLS termination, request routing by path or host, load balancing, response caching, compression, rate limiting, header rewriting, and hiding the internal topology.
A forward proxy acts for the client and is used to control outbound traffic; a reverse proxy acts for the server. Load balancers, API gateways, CDNs and ingress controllers are all specialised reverse proxies, which is why their capability lists overlap so heavily.
The architectural caution is concentration: everything passes through it, so it needs its own redundancy, and every feature added to it is a feature you can no longer test in isolation from the network.