hey team, we have a critical issue in production right now

so around 2:47 PM UTC today (Feb 12), customers started reporting they can't complete checkout. the payment form loads fine but when they hit "Pay Now" it just spins forever and then shows a generic error message "Something went wrong"

i checked the logs and there's a ton of these:
ERROR PaymentService.processPayment(): Connection timeout after 30000ms to stripe-api.internal.prod:443
the stripe internal proxy seems to be the problem. it started right after the 2:30 PM deployment (deploy #4521)

what we know so far:
- affects all payment methods (card, apple pay, google pay)
- only production, staging works fine
- the stripe dashboard shows NO incoming requests from our side after 2:30 PM
- about 340 failed transactions so far
- revenue impact estimated at $47,000
- our health check endpoint returns 200 but the payment health check returns 503

what we've tried:
- restarted the payment service pods (3 times) - no effect
- checked stripe status page - all green
- verified API keys haven't changed
- tried direct connection to stripe (bypassing proxy) from a prod pod - that works!

so the proxy is definitely the issue. but we can't just bypass it because of PCI compliance - all payment traffic must go through the proxy for logging

rollback is risky because #4521 included a database migration that added 3 new columns to the orders table

i need someone from infra to check the proxy NOW. @sarah can you look at the proxy config? also @mike can you check if the deployment changed any network policies?

current workaround: we're considering enabling the maintenance page for checkout while we fix this. estimated resolution time: unknown
