The Curious Case of the Disappearing POST 🕵️‍♂️

It started like any other release day, pipelines ran, build released, confidence high. 🚀

Then a ping from QA:

“Hey, this new API works fine in production, but not on the dev server. Getting Method Not Allowed for a POST request.”

Instant panic mode. 😨

I thought, Great, I’ve probably broken something critical… again.

First step, I double-checked the dev branch. Code was there, endpoint existed.
Then I tried the same API from my side, switched the URL to prod, and it worked perfectly.
That’s when confusion joined the party. Why on earth would the same POST fail only in dev?

A few minutes of digging later and still no clue. So I called in the big guns, @Farshan Ahmed, our Staff Engineer and Tech simbam.

He took one look at the setup and went, “Hmm…” (never a good sign).
Then, like any good detective, he pulled up Postman and examined the raw response a little closer. That’s when we saw it.
The dev server was doing something sneaky, it was returning a 301 redirect from http → https.

And here’s the kicker: a 301 redirect silently changes POST requests into GET requests.

So our poor API wasn’t broken, it was just being morphed mid-air into something it wasn’t.

No wonder the backend screamed “Method Not Allowed!”

Turns out, our nginx ingress had been using 301 for that redirect since… 2019.

Yep, this bug had been sleeping peacefully for years, until this particular POST endpoint triggered it.

The fix? Simple. We swapped out the 301 for a 308 Permanent Redirect, which keeps the original HTTP method intact.

Boom, everything worked instantly.

Crisis averted.

Feature saved.