We've got at least nine threads on outages, freezes and dropped connections, spread across the boards. They agree on almost everything, which is useful, because it means the plan below isn't one person's opinion. It's what survived contact with several bad mornings.
It's an exit problem
The outage playbook thread puts it well: a frozen platform mid-London isn't a setup problem, it's an exit problem, and you don't want to design an exit plan with your heart rate up. So write it down in advance.
Before anything goes wrong
- Stops on the broker's server, not in your head. Mental stops don't survive a disconnect. One reply in the backup power thread adds that some EAs use hidden or virtual stops that only exist inside the terminal, so check where yours actually live.
- A second path already logged in: phone app, web terminal or second device. Several threads say test it monthly, because the day you need it is the wrong day to find out the app logged you out.
- The broker's dealing desk number and your account number saved somewhere off the computer. If the machine freezes, you can't look them up on it.
- Backup internet and power. The power plan thread lists the practical options: power bank, UPS, a laptop's own battery, mobile hotspot. And test the UPS, because one with a dead battery is just a box.
The war stories thread has the lesson I'd put at the top. Failover has to be the same account on a different device, or the same device on a different network. A terminal on a backup broker does nothing for a live ticket on your main one. The platform failure thread adds a strong warning: don't open a hedge on a second device while the main one is frozen, because you can end up with a duplicate position when it reconnects.
The XAUUSD quote stall thread covers the quieter version, where gold just stops ticking for several seconds while another feed is moving. The rule there is to stop sending anything: no entries, no modifications. A modify fired at a stale quote is the one that bounces back with an error at the worst moment.
Drills, not intentions
The failover drill thread practises twice a quarter: kill the primary VPS session, confirm the secondary can flatten the same account, confirm the phone app works on mobile data, and write pass or fail with a date. "A drill I only imagine does not count" is the line from that thread worth remembering. A reply to the composite story thread suggests calling the broker's desk once on a quiet afternoon and asking how they'd close a position for you, so you know how it works before you need it.
Small disagreement
The prop phone-backup thread treats the phone as the main backup. The VPS threads prefer a second VPS session. Both work if tested. The one thing every thread rejects is the "second broker hedge" idea mid-trade.
If you only do one thing this week, log into your phone app and check your open positions and pending orders show up correctly. That's the step most people find out about too late.
What's your backup path when the main platform dies, and when did you last actually test it?
Threads this pulls from
- Broker outage playbook: how to flatten without panic
- Broker Server Downtime and Disconnections: Being Prepared
- Support quality during a platform freeze: war stories
- What to Do When Your Platform Fails Mid-Trade
- Failover: second VPS and second broker drill
- Prop scalpers: using phone as backup flatten
- The Internet Went Down With a Trade Open (Story)
- A Backup Internet and Power Plan for Traders
- XAUUSD quote stalls: what I do when the tape freezes
