MQL5 Cloud Network optimization keeps disconnecting and returning passes to queue – Build 6140

 

Hello,

I have been experiencing a serious problem with MQL5 Cloud Network optimization during the last few days.

I am a consumer of MQL5 Cloud Network computing resources, not a provider. I use the Strategy Tester to run complete optimizations through the MQL5 Cloud Network.

Until very recently, the same setup worked normally. In fact, approximately two days ago I successfully completed an optimization of more than 1,100 passes using the Cloud Network.

Now, however, optimizations repeatedly stop progressing because Cloud regions lose their connections. The affected passes are either returned to the queue or the optimization simply stops making meaningful progress.

This is not specific to one symbol or one EA. I have reproduced the problem with GOLD, BRENT and EURUSD.

My terminal is:

MetaTrader 5 x64 Build 6140
Windows 10 Build 19045

For example, during a EURUSD M15 complete optimization today:

                                                    2026.08.29 08:28:35.195 Tester complete optimization started

2026.08.29 08:33:33.583 MQL5 Cloud Europe 1: 10 gbit connection lost
2026.08.29 08:33:33.982 MQL5 Cloud Europe 8: 10 gbit connection lost
2026.08.29 08:33:34.176 MQL5 Cloud Europe 2: 10 gbit connection lost
2026.08.29 08:33:34.306 MQL5 Cloud USA 1: 10 gbit connection lost

Four different Cloud regions therefore lost their connections within less than one second.

Importantly, my broker connection did NOT disconnect at this time. The MetaTrader Journal showed no FxPro connection loss corresponding to the 08:33:33–08:33:34 Cloud disconnections.

The Cloud initialization itself also appears normal. All regions connect and authorize successfully with:

                                                    authorized (server build 6140)
common synchronization completed

In the EURUSD test, approximately 60.67 MB of history from 1971 onward was synchronized successfully in about 10.8 seconds, so the failure occurs after successful Cloud connection/authorization/history synchronization.

In other attempts I have repeatedly received messages such as:

                                                    connection lost
121 optimization passes returned to queue as not processed

followed by reconnection and authorization.

I have also seen the same general behavior with BRENT and GOLD.

I have already cleared Strategy Tester cache/logs and tester profile/state, cleared the relevant local symbol history/tick data, restarted MetaTrader/Windows and repeated the tests. The issue persists.

What makes this particularly unusual is that the same computer, same internet connection and same general optimization setup successfully completed 1,100+ Cloud passes only about two days ago.

There also appear to be recent reports concerning MQL5 Cloud agent connectivity around Builds 6090/6140, although most reports I found concern Cloud providers rather than users purchasing Cloud computing resources.

Is anyone else currently experiencing this on the consumer/Strategy Tester side of MQL5 Cloud Network?

I would especially appreciate confirmation from MetaQuotes whether there are currently any known Cloud Network connectivity/session issues affecting Build 6140.

I can provide the complete Tester and Journal logs if required.