MetaTrader Server Troubleshooting Runbook | FXSUPPORT

FXSUPPORT

Incident runbook for brokers

MetaTrader Server Troubleshooting: An Incident Runbook for MT4/MT5 Brokers

When clients report that your MT5 server is down, the first fifteen minutes decide how long the outage lasts. This runbook maps symptoms to checks and fixes across access servers, data feeds, bridges, plugins and escalation.

MT4/MT5 server troubleshooting for brokers: symptom-to-fix table, logs, return codes, access servers, data feeds, bridges, plugins and escalation steps.

By the FXSUPPORT Editorial Team · Updated 8 October 2026

Short answer: MetaTrader server troubleshooting starts with scope: is the problem one client, one group, one symbol or the whole MT4/MT5 server? Then check, in order, the server and access-server logs, connectivity, the data feed, the bridge or liquidity route, recently changed plugins and session settings. Fix the narrowest cause, communicate, and record the root cause.

Is the MT5 server down, or is it one client?

Most "MT5 server down" reports are not full outages: they are one client, one network or one symbol. Before anyone restarts a service, confirm how wide the problem is, because the wrong restart can turn a partial issue into a full one.

Each broker (or its white-label or grey-label provider) operates its own servers; MetaQuotes supplies the software:

You independently deploy our software on your own hardware, as well as maintain your server infrastructure and fully control support and administration.— MetaQuotes, MetaTrader 5 for Brokers (accessed October 2026)

On the licence page MetaQuotes adds that it has "no access to your servers, accounts, or your clients' trading history" (metatrader5.com). Traders should contact their broker. If you run a brokerage, the rest of this page is your runbook.

Use four quick scope questions:

Who? One account, one group, one IB's clients, one country, or everyone?
Where? Desktop, mobile, WebTrader, or all three? Manager and administrator terminals too?
What? Cannot log in, no quotes, frozen quotes, rejected orders, slow execution, wrong swaps or wrong session times?
Since when? What changed in the last 24 hours?

Everyone on every device suggests the trade server or the network in front of it; one region suggests an access server; one symbol suggests the data feed; orders failing while quotes flow suggests the bridge, liquidity provider or a plugin.

How the layers fit together

MetaTrader 5 spreads its work across several server roles, which is why troubleshooting means checking each layer in turn. MetaQuotes describes the platform as "based on a distributed architecture, where different functions are performed on different servers", and notes that "trade servers can be hidden behind multiple access points which will serve like shells" (MetaTrader 5 for Brokers).

 Clients (desktop / mobile / web)        Managers / admins
            |                                    |
     [ Access servers ]  <-- DDoS, routing, load
            |
     [ Trade server ] --- [ History server ] --- [ Backup server ]
       |        |   \
  plugins   data feed  gateway / bridge --> liquidity providers
 (Server API) (quotes)  (PrimeXM, oneZero, FIX ...)

Simplified MT5 incident map: check each box and each line between them.

MT4 is simpler but follows the same logic: a trade server, quote feeds, server plugins built on the Server API, and Manager API tools around it. MetaQuotes states that "the MetaTrader 4 trading platform licenses are no longer available for purchase" (metatrader4.com), so many brokers are running MT4 and MT5 side by side during a migration, which doubles the number of places to look.

The first 15 minutes: triage steps

Follow the same order every time: scope, severity, logs, recent changes, then a decision. Timestamp each step for the post-incident review.

Open an incident record with the first report time and the exact error.
Confirm scope with an internal test account on desktop, mobile and web.
Set severity using the table below, and name one incident lead who decides on restarts and client messages.
Read the logs for the last 30–60 minutes on the trade server, access servers, feed and bridge.
List recent changes: plugins, configuration, server builds, certificates, firewall, DNS, hosting.
Decide: roll back the last change, fail over, isolate a component (disable a symbol, a route or a plugin), or escalate to the vendor.
Communicate what is affected and when the next update comes.

Incident severity levels for an MT4/MT5 brokerage (example scheme)
Level	Typical situation	First action
Critical	All clients cannot log in or trade; trade server down; no prices on major symbols	All hands; fail over or roll back; status message to clients
High	One access point, one region, one platform (e.g. mobile) or one liquidity route failing	Isolate and reroute; notify dealing and support
Medium	One symbol frozen, wrong swaps, delayed reports, a single plugin misbehaving	Disable or correct the item; fix inside the business day
Low	One client's connection, cosmetic issues, warnings in logs without client impact	Ticket queue; root cause in the next maintenance window

Symptoms, checks and fixes

Match the symptom to the most likely layer, run the check, then apply the narrowest fix. Execution decisions, such as moving flow when an LP fails, belong to the dealing desk and risk policy.

MetaTrader server troubleshooting: symptom → check → fix
Symptom	Likely layer	Check	Typical fix
"No connection" for all clients	Trade server, network, access servers	Is the service running? Can an internal account connect directly and via each access point? Firewall or DNS changes?	Restart only the failed component; revert network changes; move clients to healthy access points
"No connection" for one region or ISP	Access server or routing	Connection attempts in that access server's log; latency and packet loss from the region	Take the access point out of rotation; reroute; check hosting provider status
"Invalid account" for many clients	Wrong server name in apps, group or account settings, plugin	Server names published to clients; recent group edits; plugin that blocks logins	Correct server list or group permissions; roll back the plugin
Quotes frozen on one or more symbols	Data feed	Feed status and last tick time; symbol session settings; feed source status	Restart or switch the feed; correct sessions; temporarily disable trading on the symbol
Orders rejected or "off quotes" while prices move	Bridge, gateway or liquidity provider	Bridge logs, LP session state, symbol mapping, margin at the LP	Re-establish the session; fix mapping; route to a backup LP or B-book per your risk policy
Slow execution, requotes, timeouts	Server load, bridge latency, plugin	CPU, memory and disk on the server; bridge round-trip times; plugin processing time	Remove the bottleneck; disable the slow plugin; scale resources
Server crash after an update	New build or plugin	Crash time against the change log; plugin load messages	Roll back the plugin or configuration; test in a staging server first
Wrong swaps or rollover charges	Swap settings, session or DST configuration	Swap values, triple-swap day, server time zone, rollover time	Correct and recalculate; document client adjustments
Markets open late or close early	Session and holiday schedules, DST	Symbol sessions against the venue and LP calendar	Fix sessions; publish the corrected schedule

Read the logs and return codes first

Logs turn "the server is down" into a specific component and time. Pull trade server, access server, feed and bridge logs for the same window, align time zones, and find the first error, not the loudest.

Client terminals keep evidence too. MetaTrader 5 writes platform events to the Journal tab, with a "Source" column (for example "Network") and a context-menu command to open the folder with the log files; Expert Advisor events go to the Experts tab (MetaTrader 5 Help: Platform Logs).

Codes reported by clients and Expert Advisors quickly place the fault (official MQL4 and MQL5 references):

MT4 and MT5 codes that point to server-side issues
Platform	Code	Official description	Where to look first
MT4	6 ERR_NO_CONNECTION	No connection with trade server	Trade server, access points, network
MT4	65 ERR_INVALID_ACCOUNT	Invalid account	Server name, account and group settings
MT4	136 ERR_OFF_QUOTES	Off quotes	Data feed, bridge, LP pricing
MT4	132 ERR_MARKET_CLOSED	Market is closed	Symbol sessions, holidays, DST
MT5	10031	No connection with the trade server	Trade server, access servers, network
MT5	10021	There are no quotes to process the request	Data feed, symbol settings
MT5	10018	Market is closed	Sessions, holidays, DST
MT5	10006	Request rejected	Bridge, gateway, plugin or dealer rules
MT5	10012	Request canceled by timeout	Execution path latency

If many different clients report the same code at the same minute, the cause is almost always on the server side. For a wider list, see our guide to MetaTrader common errors, and for structured log work, MetaTrader log analysis.

Access servers and connectivity

Access servers are the doors clients use to reach the trade server, so a fault there looks like "server down" even when the trade server is healthy. Check each access point separately before touching the trade server.

Reachability: test an internal login through each access point.
Load and attacks: sudden spikes in connections from unusual sources suggest a DDoS. Filter or rotate the affected access point rather than restarting the trade server. See DDoS protection.
Server lists: after changing access points, confirm the addresses clients receive. MetaTrader 4 expects the address in the format "[internet address of the server] : [port number]" (MetaTrader 4 Help).
Certificates, DNS and firewalls: expired certificates, changed DNS records and new firewall rules are common causes of "worked yesterday, fails today".
Hosting: check the provider's status page and your network monitoring. FXSUPPORT runs infrastructure in the UK and the Netherlands; see MetaTrader server hosting.

Data feed problems

If quotes stop, trading stops. MetaQuotes' own MT4 help puts it plainly:

If a client terminal does not receive quotes, it is impossible to trade with it.— MetaQuotes, MetaTrader 4 Help: Settings, Server (accessed October 2026)

Feeds reach MetaTrader through feeders and gateways: MetaQuotes describes the MT4 DataFeed API as "designed for the development of any news and quote data feeds" (MetaTrader 4 API), and the MT5 Gateway API as used for "integration with other trading systems and creation of custom data feeds" (MetaTrader 5 for Brokers). Checks, in order:

Which symbols are frozen: all, one asset class, or one feed source?
When did the last tick arrive? Does the feed log show a disconnect or login error?
Is the upstream source delivering prices elsewhere?
Do symbol names and mappings still match after a recent change on either side?
Are the symbol's sessions open? A "frozen" symbol is sometimes simply a closed session.

Fixes range from restarting the feeder to switching to a backup source. While prices are stale, consider disabling trading on affected symbols. More detail: MetaTrader data feed setup.

Fighting the same incident every month?

Send us the symptom, the time it happens and your recent change log. We will tell you which layer we would check first and how we would stop it recurring.

Request a free technical review WhatsApp Telegram

Bridge and liquidity issues

When quotes flow but orders fail, look at the bridge or gateway between MetaTrader and your liquidity providers. Bridges such as PrimeXM XCore connect MT4/MT5 to external liquidity; PrimeXM describes XCore as supporting "multiple industry standard interfaces – such as MT4/MT5 and FIX API protocols" (PrimeXM), and oneZero describes its Hub as a "pricing & risk management system" (oneZero).

Session state: are the FIX or API sessions to each LP logged in? Look for disconnects, sequence-number resets and heartbeat timeouts.
Rejections: read the reject reason. "Off quotes", price-tolerance and volume rejects point to different fixes.
Mapping: after symbol changes, check that MetaTrader symbols, contract sizes and digits still map correctly to LP instruments.
Credit at the LP: rejects can mean your LP account is near its limit.
Latency: compare bridge round-trip times with your normal baseline; see latency optimisation.

Bridge and LP faults usually need the vendor. FXSUPPORT works with technologies including PrimeXM XCore, oneZero Hub, Match-Trade and FIX API; see bridge solutions, PrimeXM XCore integration and oneZero Hub integration.

Plugin conflicts and update problems

Server plugins run inside the trade path, so a faulty plugin can block logins, slow execution or crash the server. MetaQuotes notes that the MT4 Server API allows plugins covering "management of server parameters, order and customer base, processing trade requests and more" (MetaTrader 4 API).

Compare the incident start time with the plugin change log and server restarts.
Check the server log for plugin load errors and slow processing.
Look for two plugins acting on the same event.
Test disabling the suspect plugin on staging first; on production, only with the incident lead's approval.
After a server build update, confirm every plugin is compatible before reopening trading.

The long-term fix is change control: a staging server, a written rollback plan and one owner per plugin. We build and maintain plugins; see MetaTrader plugin development.

Sessions, holidays and daylight saving time

Many "Monday morning" incidents are schedule errors rather than technical faults. Wrong session times close markets early, delay the open or charge swaps on the wrong day.

Daylight saving time is the classic trap because regions change on different dates. In 2026 the UK clocks go back on 25 October (GOV.UK), while US daylight saving time ends on 1 November (NIST). For that week, US-linked sessions sit one hour away from their usual position relative to London. Check sessions, swap times and your LP's schedule before each change and before holidays. Session and DST/holiday schedule checks are part of FXSUPPORT's risk operations service; see risk management.

Server resources, databases and backups

Slow degradation, such as execution times that grow every week, usually means a resource limit. Watch CPU, memory, disk space and I/O on every server and alert before thresholds are reached.

Disk: logs and history grow continuously. A full disk can stop logging, reports and backups at the same time.
Backups: confirm that backups complete and that you have tested a restore. MetaQuotes notes that MT5 system databases come with "backup and recovery options" (MetaTrader 5 for Brokers); a backup you have never restored is still an assumption.

We support both Windows Server environments, typical of MetaTrader, and Linux environments used for web, database and supporting systems. See server monitoring, backup and recovery and performance tuning.

Escalation: who owns which fix

Escalate by layer: each component has an owner, and knowing who to call in advance saves the most time.

Escalation matrix for an MT4/MT5 incident
Layer	Owner	Bring to the call
Trade, history and access servers	Broker's MetaTrader admins or support provider (grey-label clients: the server licence holder)	Logs, times, recent changes
MetaTrader software defects	MetaQuotes, via the licence holder's support channel	Build number, reproduction steps, logs
Bridge or gateway	Bridge vendor (e.g. PrimeXM, oneZero, Match-Trade)	Order IDs, reject reasons, session logs
Liquidity and pricing	Liquidity provider	Symbols, times, price samples, rejects
Network, servers, DDoS	Hosting provider or data centre	Traceroutes, traffic graphs, affected IPs
Plugins, CRM, API integrations	Plugin developer or CRM vendor	Plugin version, change log, API errors

A grey-label broker, set up inside another company's licensed MetaTrader 5 server with its own client groups and brand while the MT5 server licence stays with the provider, usually cannot fix server-level faults itself and must escalate to the provider. See white and grey label.

After the incident: review and regulators

An incident is closed when the root cause is known and a preventive action agreed, not when the server is back up. Write a short report: timeline, impact, root cause and preventive actions with owners.

The FCA notes that firms in scope of its operational resilience rules "had until 31 March 2025 to ensure they could operate their important business services within their impact tolerances" (FCA). In the EU, ESMA states that the Digital Operational Resilience Act (DORA) "entered into force on 16 January 2023 and will apply as of 17 January 2025" (ESMA). Whether and how you must report a given incident depends on your licence and jurisdiction; check with your compliance officer. This is general information, not legal advice.

How FXSUPPORT helps

FXSUPPORT provides outsourced MT4/MT5 technical support and managed operations for forex brokers and prop firms, including incident troubleshooting. FXSUPPORT is not affiliated with MetaQuotes or with any vendor named on this page.

Platform support

Setup, updates, backups, configuration, log monitoring and bridge configuration for MT4 and MT5. See MetaTrader support.

Infrastructure

Servers in the UK and the Netherlands, monitoring, DDoS and security hardening, with a 99.9% uptime target.

Risk operations

End-of-day and evening reports, toxic-flow lists, and session and DST/holiday schedule checks.

Root-cause reports

After significant incidents: timeline, root cause, impact and recommended preventive measures, plus disaster recovery planning and restores.

Pricing depends on your platform (MT4, MT5 or both), number of servers, integrations, coverage hours and whether you need one-off help or ongoing operations. Contact us for a quote.

Frequently asked questions

Is MT5 down for everyone?

There is no single MT5 server. Each broker runs its own servers, so an outage at one broker does not affect others. Traders should contact their broker; brokers should confirm scope with an internal test account.

Can MetaQuotes fix my MT5 server when it is down?

MetaQuotes says it has no access to brokers' servers, accounts or trading history, and that brokers maintain their own infrastructure. MetaQuotes can help with software defects through the licence holder's channel, but day-to-day operation and incident fixing are the broker's or its support provider's job.

Why are clients getting "No connection" when the server is running?

Usually an access server, route, DNS record, certificate or firewall rule is failing in front of a healthy trade server. Test each access point before restarting anything.

Why do many clients suddenly see "Invalid account"?

Common causes are a wrong server name in the server list clients receive, a group or account setting change, or a plugin that blocks logins. Compare the start time with your change log.

Quotes are frozen on one symbol. Where do I start?

Check the symbol's session settings first, then the last tick time and feed log, then the upstream source and symbol mapping. Consider disabling trading on the symbol while prices are stale.

Orders are rejected but prices are moving. What is wrong?

Look at the execution path: bridge or gateway sessions, reject reasons, symbol mapping and your margin or credit at the liquidity provider.

How do I know if a plugin is causing the problem?

Check whether it started after a plugin change, look for plugin errors in the server log, and test by disabling the plugin on staging first.

Why do problems often appear around daylight saving changes?

Regions change clocks on different dates. In 2026 the UK changes on 25 October and the US on 1 November, so sessions, rollover and swap times can be off by an hour if they are not updated.

Do I have to report a platform outage to my regulator?

It depends on your licence and jurisdiction. UK firms in scope of FCA operational resilience rules and EU firms in scope of DORA have specific obligations. Check with your compliance officer; this is general information, not legal advice.

I run a grey-label MT5 brokerage. Can I fix server issues myself?

Usually not at server level, because the MT5 server licence stays with the provider. Server, access server and feed problems are escalated to the provider.

Is MT4 still supported for troubleshooting?

MetaQuotes no longer sells MT4 licences, but many brokers still run MT4 servers. We support MT4, MT5 and mixed setups, including MT4 to MT5 migration.

How much does incident support cost?

It depends on your platforms, number of servers, integrations, coverage hours and whether you need one-off help or ongoing operations. Contact us with your setup for a quote.

Sources

MetaTrader 5 for Brokers — MetaQuotes (accessed 8 October 2026)
Buy the MetaTrader 5 Platform — MetaQuotes (accessed 8 October 2026)
MetaTrader 4 for Brokers — MetaQuotes (accessed 8 October 2026)
MetaTrader 4 API — MetaQuotes (accessed 8 October 2026)
MetaTrader 4 Help: Settings, Server — MetaQuotes (accessed 8 October 2026)
MetaTrader 5 Help: Platform Logs — MetaQuotes (accessed 8 October 2026)
Trade Server Return Codes — MQL5 Reference (accessed 8 October 2026)
Error Codes — MQL4 Reference (accessed 8 October 2026)
XCore — PrimeXM (accessed 8 October 2026)
oneZero trading technology — oneZero (accessed 8 October 2026)
Operational resilience — Financial Conduct Authority (accessed 8 October 2026)
Digital Operational Resilience Act (DORA) — ESMA (accessed 8 October 2026)
When do the clocks change? — GOV.UK (accessed 8 October 2026)
Daylight Saving Time — NIST (accessed 8 October 2026)

FXSUPPORT is not affiliated with MetaQuotes Ltd. or the vendors named on this page. General information, not legal or investment advice.

Need a second pair of eyes on a live incident or a recurring fault?

Send us the symptoms, recent changes and log extracts. We review your MT4/MT5 setup and tell you what we would check and fix first.

Request a free technical review View All Services

Explore More