Also rumor has it they “lost user data and the backups are not working as expected”
Edit: This is the tweet. Not sure if it is legit: https://mobile.twitter.com/classrobinhood/status/12345774264...
The question is really does the front-end normally try to query the next day and what's going on in their order execution system on the backend, which we don't know because this source is just speculation based on console logs in the browser.
Additionally, if you’re a conspiracy theorist — and who isn't, in this wild 2020 ride — Robinhood was down exactly four years ago today, too, lol
I didn't even know they had a phone number you could call. This should have been emailed out after the first hour of the outage. Or, how about you have a better system down page then just a blank website? If API health checks fail, serve a static site from S3 (route53 can do this). The ops failure of a company that can afford to pay quality DevOps engineers $200-300k a year is beyond belief.
Simple as:
"We are experiencing a technical outage... <corporate blah blah blah here>. If you need to make trades, please do so via phone at 1-800-SHORT-RH"
"You're call will be taken in the order it was rexeived. Current wait time: 1,732 minutes."
Robinhood would be using a proper date library anyway to handle timezones. The system was fine on the 29th, 1st and even this morning pre-market. Seems more like a capacity issue that cascaded into failure.
My bet is still on it being a database issue. If they indeed are using AWS RDS, I wouldn’t be surprised if they have half of AWS support engineers working on it. Once again, if I had to bet, I say they rolled their own "half-baked" database solution instead of using a AWS native offering.
(yes: the URI says February 5th whilst the article says March 5th!)
A company like Robinhood doesn't strike me as a strong candidate for contracting AWS Enterprise support. For the price you get almost nothing out of it.