Ask HN: What services do you use to send email for small scale apps?
Services like Sendgrid and Mailgun offer free tiers up to a certain number of emails sent per month (about 10k). But beyond the free allotment, the cost of sending emails begins to add up, especially when running an app that doesn't generate revenue.
How does the use of these, or similar, email services compare to using Gmail or Fastmail's SMTP servers to send all your app's emails, assuming they're mostly transactional?
What other options do you recommend for sending transactional emails at small scale (0 - 50k / month)? Amazon's email service is extremely cheap [1].
I don't think you can go cheaper than them? Do you know how SES stacks up against services that focus solely on email? It looks like a great deal if you already use EC2 because of the free 62k outgoing emails/month. But there doesn't seem to be a free tier otherwise. I like MailGun (10k/mo free) - if cost is a significant issue for larger volume why not do the work to send directly from your server? Everything I've read about hosting one's own mail servers indicates, other than the educational rewards, it's not a practical idea. But along those lines, I was considering sending emails using Gmail or Fastmail's servers. I'm unfamiliar with the potential pitfalls, if any, when going this route. I'm hoping someone can comment on that or share their experience. When deciding on architecture, I always start with asking what are my key priorities with this project? Among bulk send providers, Amazon SES is probably cheapest (I read their chart as $100 for 1 Million emails). Finally, I got something very different out of my reading on running your own mail servers. HN search (https://hn.algolia.com/?query=mail%20server&prefix&dateRange...)
I have used Gmail for little tiny projects like a contact form that got used a dozen times per year, but that was before MG was an option. As for using Gmail or FastMail today, it will probably work for now, but will it continue doing so in the future? Maybe, but my own personal experience indicates that _all_ free services are not something one can count on persisting over the long haul. TANSTAAFL = you get what you pay for. At some point their funding will run out, or new management will decide to optimize their business expenses. If I were to plan on using a service like this (IANAL) the first thing I check is the Terms of Service. The second thing I check is the service providers intended use case. (When I read the terms of use for both Gmail and Fastmail I get the distinct impression they are intended for use by a human, not for volume transactional send). While these providers may not currently prevent their service being used for transactional mail send - there is a long history of service providers throttling, ejecting, or (if paid) increasing the billing for their highest volume users. When that inevitably (considering history) happens, a new solution will have to be found (probably with minimal notice and urgent deadlines). I don't know about you but I have enough on my plate maintaining the evolving decay of my projects (as libraries fall out of favor and security updates stop) without having to worry about monitoring changes in QOS for email deliver-ability for a cobbled together solution. to minimize cost = run my own mail server
to maximize learning = write my own mail server
to maximize flexibility or do something truly unusual = one of the above
to minimize time = pay for a service
to maximize maintenance cost = cobble solution from pieces intended for other uses.