Over lunch, a friend yesterday told me a sad story about it taking them 6 months of debugging to find out why their transactional mail was bouncing. Their Open Source Eventbrite alternative, EventYay, manages event tickets for thousands of attendees each month.
They used Sendgrid to send notifications etc, but Cloudflare (SendGrid partner) had made a DNS error which caused RDNS lookups to fail for the SendGrid owned domain / IP. It took all those months for the customer to figure this out, because engineers at all other companies concerned couldn’t understand or resolve the issue. Eventually he figured it out with help from someone at Gmail (one of many destinations that mail wasn’t being delivered to).
This fascinates me: even at a premium email SaaS provider like SendGrid, basic DNS config issues can arise and go unchecked for months, costing the customer the one thing that such platforms are supposed to provide: convenience.
Perhaps there’s a need for tools, e.g. in Lightmeter, for troubleshooting the config of third party mail networks, to assist in such cases. After all, plenty of companies use a combination of self-hosted and external mail hosting (according to our latest research around 1/3 of all self-hosted organisations fall into this hybrid category).