Sending works, nothing comes in, and no bounce shows up anywhere. That asymmetry tells you almost everything: inbound mail is decided by MX records, not by SPF. This guide walks the inbound path in the order a sending server actually walks it, groups the real causes into five DNS failure families, and gives you the dig commands to prove which one you are looking at.
A limit_req block that parses cleanly can still be doing absolutely nothing, or throttling every visitor behind one CDN edge. This guide covers Nginx rate limiting and bot mitigation as they actually behave in production: getting the key right behind a proxy, what burst and nodelay do to your traffic shape, sizing zones correctly, rolling out in dry run mode, and layering bot handling from robots.txt through user agent maps to firewall escalation.
Choosing between an Application Load Balancer and API Gateway is not a feature comparison. It is a choice of hard limits, a billing curve and an authentication story. Here is how each one bills you, which constraints surface months later, what ALB's native JWT validation changes, and a decision procedure you can run in ten minutes.
Your domain has all four records. Every free checker shows green. It is still trivially spoofable, because publishing a record and enforcing a policy are different acts. A working guide to business email DNS: what MX, SPF, DKIM and DMARC each actually protect, the SPF lookup budget and the void-lookup limit nobody watches, why DKIM alignment decides whether you can ever reach p=reject, the domain architecture that stops the problem recurring, and a rollout order that gets you to enforcement without losing legitimate mail.
Most "the site is down" tickets are really DNS problems wearing a costume. Here's the practical DNS troubleshooting workflow I use — how to read dig output, follow the delegation chain, decode SERVFAIL and NXDOMAIN, and stop blaming propagation for things that are actually caching, TTL, or DNSSEC failures.