Load balancers

How this question breaks down

1. 1BasicsCore8 minSpread requests across a pool of servers, notice the ones that fail, and keep the balancer itself from becoming the single point of failure.
2. 2L4 vs L7Core22 minA layer 4 balancer picks a backend once per connection from addresses and ports; a layer 7 balancer reads every request and can route, retry and rebalance per request, at the price of terminating TLS and parsing HTTP.
3. 3AlgorithmsCore18 minHow a balancer picks the next backend, from round robin that knows nothing to least-request and hashing that know load or keys, and why the best choice depends on how uneven your requests are.
4. 4Health checksOperations21 minHow a balancer learns that a server is sick (probes and real errors), how fast it can know, what it does when too many look sick, and how to take a server out without dropping a request.
5. 5GlobalOperations18 minChoosing which region serves a user, from where the user is, which regions are healthy and which have room, and handing that choice to the local balancer inside the region.

Work through the parts in order and you’ve covered the full design. Each part is one topic.