Tips for reducing API requests and bandwidth usage on a high-traffic Next.js blog

Hi, I have a next.js blog that receives a lot of traffic. Currently, my API Requests and Bandwidth usage are 30x more than what's included....with very expensive overage costs 😞 Is there anything I can do to reduce this usage? Any guides to refer to I'm not finding? I won't be able to afford this long term. Thank you

Are you using getStaticProps for fetching blog data or getServerSideProps?

GetStaticProps

Do you host any videos or high-resolution images on your blog? Also, DM me your project ID!

GetStaticProps

My requests and bandwidth were also quite more than expected; in my case it had to do with non-human bots visiting the site, which triggered the code of the page just like a regular visit, which means my queries ran.
I was in a position where I could safely keep them from visiting without adversely affecting the site. If you have or can create/track some sort of log mechanism, pay attention to where the visits are coming from and, if possible, the user agents. There may be an opportunity to trim the fat ( just make sure you still allow responsible/moderate and probably-necessary bots like Google to visit )

We were headed for 25/mo between bandwidth and requests and now it's probably going to be 2 to 4.

Thank you for the tip. Will look into bots -- had a logger running on my Vercel instance for the past week. Doing a deep dive today. How did you engineer the bot wall?

We are actually using the PHP version of the Sanity client and on and old-timey Apache site, but one with a weird config and that we couldn't use .htaccess or robots.txt for; the client was imported with the require function into the few relevant pages that needed to pull in Sanity content, so I used PHP to look at the user agent string and wrapped the client's "fetch" command in a condition that only let the code inside get run if it it didn't match one of a few different kinds of substrings ( like "petal" for PetalBot, for example). They still visited but they stopped triggering my Sanity code.
In the case of the particular bots we targeted, like an uptime bot, no issue arrises from that content missing from the page, so it was okay to proceed with that approach.

If I ever do something Next-y or
jamstack -y, I might try the FingerprintJS middleware

Wow cool. Super helpful, thank you!