MT4 EA redundancy solution?

Imho it doesn't make much sense to invest time in investigating closed-source related issues. For a matter of fact an experienced trader will not run a closed-source EA on a live account, so the questions and ideas are purely theoretical.

Instead I would focus on the real issue. Why do you have the need for redundancy in the first place? Obviously, as you stated, your server is of low quality. So, this is the issue to fix, and the issue can be fixed easily. For less than $50 you get 99% uptime garanty, so this should be fixed first. Setting up two low-quality servers trying to watch each other will not improve your situation but only cause sleepless nights. Do setup one high-reliable server, and setup bullet-proof monitoring and signaling for this one. That gives you an idea of how many outages really occur, and gives you a way to manually intervene. And manually intervene you will have to, that's for sure.

Manage your stops invisibly, and use visible desaster stops. That gives you a time buffer for manual intervention and the real safety net.

I am using a VPS from a company called NY City Servers which have a 100% Uptime Guarantee. The concern I have is that the terminal or windows crashes and then the MT4 terminal has to be reset/restarted. Even though the server is up 100% the software like Windows or MT4 doesn't necessarily remain up 100%.

Even though I don't have the EA source code I know how it functions and its strategy, so I'm comfortable to use it live.

Maybe this is the .challenge??

Let us use our imagination.....

When I say heart beat I am referring to the main mt4 terminal connectiviity which of course you would produce with an EA or indicator.
The no source EA would only serve to make the trades which will be monitored and logged by the Redundant EA.

I would love to take the challenge but I am already quite busy with 3 other EAs.

Whenever times permit I will get some of my code and put together a simple client/server indicator to illustrate the basic operations Iam eluding to. It will show how it can detect the "heartbeat" of another one or several terminals and it will illustrate how the redundant indicator records the info from the other terminal so that it knows the state of the main terminal or terminals.



I

If the terminal crashes: You can monitor a terminal from another terminal (or another application) on the same machine
If Windows crashes: That doesn't happen and you don't need to prepare for it. If you want you can setup a cheap Linux box with Nagios and Nagios web interface. Then you monitor the login port. For login you use SSH and tunnel everything else (RDP) through it. Your SSH port is the only open port.

From 5 years experience with Forex VPS I can tell you that the important issue is not "crashing". In 5 years a terminal crashed 1 or 2 times, and I usually run 3 to 5 in parallel without issues. What will happen is:

1. The terminal loses network connectivity because of a glitch (but it doesn't crash) and it takes up to 5 minutes (typically about 10 seconds) to automatically reconnect. Check your logs and you will see how often/long it happens. Block outgoing traffic from your terminal and allow only traffic to your broker to prevent MetaQuotes issues with Market Place crap and so on. Can be monitored from the same machine.
2. Your broker's trade server goes down (that's the remote data center). Happens about once a year and can be a serious issue if it lasts. You need to prepare for that case, as you have to call-in and to close your trades manually like we did in the old days. Nothing can be done from a second machine/EA.
3. You own network link goes down (that's the local data center). Happens rarely but still happens. Again can be monitored from the same machine.
4. 95% of the issues I have are caused by bugs in my code. For closed-source this translates to bugs in the strategy. Little can be done from a second machine/EA.

In summa: The issue is not crashing applications or servers but network and strategy. But still the ZeroMQ recommendation is good advice. If you have the skills/resources you can setup a two-way communication channel to your EA. I prefer the more simple IRC as free client applications are available for every device. From the mobile you can send a command to the EA and it responds back to the mobile e.g. with it's status. Of course this requires your own IRC server, you don't want to publish live account data to a free/public IRC server.

ps: In case your VPS goes down unannounced or in RTH/ETH you know that the advertized 100% are purely a marketing gag. I don't know any professional data center offering 100% uptime garanty to any customer in the universe. In IT that's simply not possible. My advice for you would be (if I may) to start *there*.

Thank you all for your time. The strategy I have can potentially have trades managed for a long time (over 12 months) which relies on the cached memory so a terminal crash (even though rare) would mess things up considerably. I will keep looking for a solution to deal with the rare terminal crash scenario and how to log the memory cache for MT4.

As already said, immediately log each state changing event to a resource outside of the terminal. On restart restore state from there. This is not complicated and handles crashes perfectly. It can be a file in the presets folder.