How to Monitor Multiple Ignition Gateways: Tools and Setup
How to monitor the health of multiple Ignition gateways in one place: EAM, System tags and alarms, Catalyst in Ignition 2027, and analyzer modules compared, with documented thresholds, alert setup steps and a root cause checklist.

To monitor the health of multiple Ignition gateways in one place, connect them over the Gateway Network and make one gateway the controller of the Enterprise Administration Module (EAM). The controller gets System tags for every agent gateway and raises Agent Events and alarms when CPU, memory, error rates or database pool use cross its thresholds. For the root cause of one slow gateway, add a per gateway diagnostic layer: the Diagnostics pages, Catalyst in Ignition 2027, or an analyzer module.
What is the best tool to monitor multiple Ignition gateways?
No single tool does everything well. The honest answer depends on the job:
- Fleet alerting across many gateways: EAM. It is the only built in option designed to watch many gateways from one central gateway and alarm on their health.
- A few gateways without EAM: alarms on System tags on each gateway, plus remote tag providers so one gateway can display the others' status.
- Root cause on one gateway: the gateway's own Diagnostics pages, the Catalyst Diagnostic Assistant on Ignition 2027, or a dedicated analyzer module that scores and reports findings.
Most teams with more than a handful of gateways combine a fleet layer that tells you which gateway is unhealthy with a per gateway layer that tells you why.
Figure 1. A common layout: every site gateway is an EAM agent, one controller watches them all and alarms through a pipeline, and deep diagnostics stay on each gateway. Component names follow the Inductive Automation documentation.
Built in option 1: the gateway Diagnostics and status pages
Every Ignition gateway ships with its own diagnostic web pages. In Ignition 8.3 they live under Diagnostics; in 8.1 the equivalent pages sit in the Status section. According to the Inductive Automation documentation, the 8.3 Diagnostics area includes:
- Overview: version, license state, uptime, connections outlined in red when they have issues, memory, CPU and environment details.
- Metrics Dashboard: user configurable dashboards of gauges, meters and other metrics covering performance, tags, alarming, projects and databases.
- Logs: errors from connections, authentication, alarm journals and pipelines.
- Execution: scheduled task status, including how long a tag group takes to execute.
- Performance: current CPU, memory in use against the total allocated, thread performance, response time, CPU and memory trends for the previous 24 hours, and recent slow response events.
- Scripts: running and configured gateway scripts, with the option to terminate a running script.
- Threads: what each thread is doing, with its state and CPU use.
These pages are the best place to investigate, but they are per gateway: ten gateways means ten browser tabs and nothing telling you which to open first.
Built in option 2: the Gateway Network with EAM
The documentation describes EAM as a module that lets you control and monitor any number of Ignition gateways from a central gateway, using the Gateway Network to move data between them. One gateway is the controller; every other gateway on that Gateway Network is an agent. Only one controller is allowed per Gateway Network.
What the controller watches
Every agent reports its general health to the controller. When something goes wrong, the controller records an Agent Event. The documented examples are high CPU on an agent, a large number of system errors in a short time, unusually high use of database connections, and loss of connectivity to the agent. Agent Events are stored in the controller's database in an agent_events table, and the controller can turn them into alarms and alarm notifications.
The controller also creates System tags for each agent, which is what makes a single monitoring screen possible. Per the System tags reference, each agent folder contains:
- Root tags such as IsConnected, IsRunning, LastCommunication, RunningState (Disconnected, Running, Warned or Errored), Version and AgentGroup.
- Metrics / System: CPU Usage and Memory Usage, where memory is reported in whole percentages from 0 to 100.
- Metrics / Database: one folder per database connection with Active Connections, Active Queries and Active Query Duration.
- Metrics / Logging: Errors Per Minute and Errors Per Hour.
- Metrics / Session: Vision clients, Designers and Perspective sessions.
- ModuleVersions: one tag per installed module with its version.
Because these are ordinary tags on the controller, you can alarm on them, trend them and build a single overview screen that covers every gateway.
What else EAM does for a fleet
The controller can also schedule tasks against agents, such as collecting backups, installing modules, remote upgrades and sending projects. On large networks, Inductive Automation suggests adjusting the Send Stats Interval in Agent Settings so agents send statistics less often.
Without EAM: Gateway Network tags and remote tag providers
If you are not running EAM, the Gateway Network still helps. Each gateway gets System tags for every remote gateway it is connected to, including IsAvailable and LastComm. A Remote Tag Provider lets one gateway read another gateway's tags over the Gateway Network, read only by default. Combine the two and a central gateway can display the System tags of the others, though alarm evaluation still happens on each remote gateway.
Built in option 3: System tags and alarms on each gateway
Each gateway exposes its own health in the System tag provider. System tags cannot be deleted or modified, but the documentation confirms you can configure alarms on them. The most useful folders for health monitoring are:
- Performance: CPU Usage, Memory Usage (bytes), Max Memory (bytes), Memory Utilization (current usage divided by maximum), Disk Utilization and Available Disk Space (MB).
- Database, one subfolder per connection: Available, AvailableThroughFailover, ActiveConnections, ConnectionSaturation (active connections against maximum possible), AvgQueryTime and QueriesPerSecond.
- Store and Forward, per connection: TotalPendingCount, TotalDroppedCount and TotalQuarantinedCount.
This works on every gateway with no extra modules. Its weakness is scale: every alarm lives on its own gateway. The setup section below shows how to copy alarms between gateways with a script.
Is Catalyst in Ignition 2027 enough for gateway diagnostics?
Ignition 2027 introduced Ignition Catalyst, an AI framework built into the platform across two modules, Catalyst Core and Catalyst Agent. Here is what the official Inductive Automation sources say about its diagnostic side:
- Catalyst Core is bundled with the Ignition platform and adds Gateway Diagnostic Assistants to gateway web pages.
- Diagnostic Assistants are currently available on the Perspective Sessions, Store and Forward, Database Connections, OPC UA Connections, Tag Providers, Event Streams, Running Scripts, Gateway Scripts and Threads pages.
- The assistant can run automatically when a page loads, on demand with a button, or be disabled. A Run deeper diagnostic button produces an in depth summary of the issues on that subsystem and suggests potential fixes.
- You must connect a model provider first. The documented provider types are Anthropic, Ollama, OpenAI (or OpenAI compatible endpoints) and OpenRouter. Inductive Automation's announcement points to self hosted options like Ollama for air gapped, on premise deployments.
- Inductive Automation announced Ignition 2027 as a public beta, with general availability targeted for February 2027.
So is it enough? For explaining what is wrong on one page of one gateway, Catalyst looks useful, and it is included with the platform. It does not replace a monitoring setup, for two reasons that follow from the documentation:
- It is per page and per gateway. It explains the page you are looking at. It is not documented as a fleet view, a threshold engine or an alerting system.
- It depends on an AI model. Sites that cannot send gateway data to a cloud provider need a self hosted model, and someone has to own that.
The sensible pattern is the same as before: EAM or System tag alarms tell you which gateway needs attention, then Catalyst, the Diagnostics pages or an analyzer module help you find out why.
What to monitor: the key gateway health signals
The only health thresholds Inductive Automation publishes as numbers are the defaults on the EAM Event Thresholds page. Use them as a starting point. For everything else, set a threshold that fits your baseline: watch the value for a few normal weeks, then alarm on what is clearly abnormal for that server.
- Memory: EAM defaults are a warning at 70% and an error at 90% of the agent's RAM usage. On a single gateway, watch Memory Utilization and Memory Usage against Max Memory.
- CPU: EAM defaults are 70% for warning and 90% for error.
- Database connection pool: EAM defaults are 80% utilization for warning and 100% for error. On a single gateway, ConnectionSaturation is the matching tag.
- Database availability: the Database Available tag is a true or false value, so alarm as soon as it reads false. If you use failover, also watch AvailableThroughFailover.
- Query time: AvgQueryTime has no documented threshold. Set one that fits your baseline.
- Error rate: EAM defaults are 2 and 5 errors per minute, and 20 and 60 errors per hour, for warning and error.
- Agent inactivity: EAM raises a warning after 5 minutes without contact and an error after 15 minutes by default.
- Sessions and clients: EAM defaults are 50 and 100 for connected clients and for Perspective sessions.
- Store and forward: no documented threshold. Watch TotalPendingCount against your baseline and investigate any growth in TotalDroppedCount or TotalQuarantinedCount.
- Disk: no documented threshold. Set one that fits your baseline and leaves room for logs and backups.
- Gateway Network links: alarm when IsAvailable on a remote gateway reads false.
Figure 2. The default warning and error levels on the EAM Event Thresholds page in the Ignition 8.3 documentation. Treat them as starting points and tune them to your own baseline.
Step by step: alerts for low memory and lost database connections
This setup needs no extra modules. Repeat it per gateway, or use the EAM version below to cover the fleet from one controller.
Before you start
Alarm notification pipelines need at least one alarm notification profile and at least one on call roster configured in the gateway, and each user in the roster needs valid contact details such as an email address. Set those up first.
Single gateway setup
- Create the pipeline. In the Designer, open the Global node of the Project Browser and create an alarm notification pipeline. The documentation's example of a Delay block before a Notification block, with Cleared as a dropout condition, means only alarms active longer than the delay send a notification.
- Open the System provider. In the Tag Browser, choose the System tag provider and expand Gateway.
- Add the memory alarm. Open [System]Gateway/Performance/Memory Utilization in the Tag Editor, select Alarms and add an alarm with the Above Setpoint mode. Check the live value in the Tag Browser first so you know whether your version reports it as a fraction or a percentage, then set the setpoint to match your baseline. The EAM default of 70% for warning is a reasonable reference.
- Add the database alarm. Open [System]Gateway/Database/<connection name>/Available, where the folder name is the exact name of your database connection, and add an alarm that triggers when the value is false, for example the Equal mode with a setpoint of 0. Give it a high or critical priority.
- Add an early warning. On the same connection, add an Above Setpoint alarm on ConnectionSaturation. The pool filling up often comes before failed queries.
- Route the alarms. Set the Active Pipeline on each alarm to the pipeline you created.
- Test it. On a test gateway, disable the database connection and confirm the alarm goes active and the notification arrives. Re enable it and confirm the alarm clears.
- Copy it to other gateways. The Tag Browser cannot export System tags, so the documentation uses a script: run system.tag.exportTags(filePath, ["[System]"]) in the Script Console, then import the JSON file on the next gateway. Importing overwrites matching System tag settings.
Exact folder names can differ between Ignition versions, so confirm each path in your Tag Browser before you build alarms on it.
Fleet setup with EAM
- On the controller, go to the Network page and select Enterprise > Event Thresholds.
- Enable Activity Alarms for agent inactivity and Metrics Alarms for CPU, memory, error rate and database utilization. Enable Task Alarms if you also want to know when a scheduled task fails.
- Set the Warning Priority and Error Priority.
- Set the Active Pipeline using the documented format project:ProjectName:/pipeline:PipelinePath. The pipeline must exist before an event happens, and the name is case sensitive.
- Adjust the memory, CPU and database utilization thresholds, which default to 70 and 90, 70 and 90, and 80 and 100 percent.
Note that EAM's memory metric is a percentage from 0 to 100, while the local Performance tags report bytes and a utilization ratio.
Root causing a slow Ignition gateway: a checklist
When a gateway feels slow, work from the outside in. This checklist uses the pages and tags Inductive Automation documents for 8.3.
- Confirm it is the gateway. Check the Performance page for CPU, memory against the allocated maximum, response time and slow response events, and use the 24 hour trends to time the slowdown.
- Look at threads. The Threads page shows what each thread is doing and its CPU use. One busy thread usually points to a specific script, query or driver.
- Check running scripts. The Running Scripts page lists active gateway scripts and lets you stop one that is stuck.
- Check scheduled work. The Execution page shows durations for scheduled tasks, including how long tag groups take to execute.
- Check the databases. Compare AvgQueryTime, QueriesPerSecond and ConnectionSaturation against normal. A saturated pool makes everything that touches the database wait.
- Check store and forward. Rising TotalPendingCount or any quarantined data means data is backing up between the gateway and the database.
- Check the Gateway Network. The Gateway Network Performance page shows queues and thread pools. Inductive Automation notes that thread pools default to a limit of 5 requests, and gives an example where slow tag history queries over the network delay other traffic such as remote tags.
- Read the logs. Database, device, authentication and pipeline errors all land in Logs.
- Capture a diagnostics bundle. It includes a thread dump, wrapper logs and metrics data, plus a memory dump if enabled and free disk is at least twice the max heap. The documentation warns that large bundles can affect performance or restart the gateway.
- Ask for a summary. On Ignition 2027, run the Catalyst deeper diagnostic on the relevant page. On 8.3.2 or newer, an analyzer module can rank findings across these areas at once.
Ignition gateway monitoring tools compared by use case
Built in Diagnostics and status pages
- Best for: investigating one gateway in depth.
- Strengths: included with every gateway, detailed thread, script, database and performance views.
- Limits: one gateway at a time, no alerting on their own.
Enterprise Administration Module (EAM)
- Best for: watching and alarming on many gateways from one controller.
- Strengths: per agent System tags, documented default thresholds, Agent Events stored in a database, alarm pipelines, plus backups, upgrades and project distribution.
- Limits: it tells you which gateway and which metric, not the detailed cause. It is a separate module to license and set up.
System tags with alarms
- Best for: small sites or teams not ready for EAM.
- Strengths: no extra modules, flexible alarm modes, works on every gateway.
- Limits: configured per gateway, so it takes discipline to keep alarms consistent across a fleet.
Catalyst Diagnostic Assistant (Ignition 2027)
- Best for: plain language explanations of what is wrong on a specific gateway page.
- Strengths: bundled with the platform, suggests potential fixes, can use a self hosted model.
- Limits: needs Ignition 2027 and a configured AI model provider, works page by page, not a fleet or alerting tool.
Plantformance Gateway Analyzer
- Best for: a structured health review of a gateway, with scored findings and a report you can share.
- Strengths: one health score, findings grouped by severity, JSON and CSV exports, scheduled report delivery and a diagnostic tag provider.
- Limits: runs on and analyzes the local gateway, requires Ignition 8.3.2 or newer, and is not a central multi gateway product.
Other documented options
- Metrics Dashboard: custom gauges, meters and histograms per gateway in 8.3.
- Remote tag providers: a central Perspective or Vision screen that reads other gateways' System tags over the Gateway Network.
Criteria to choose
- How many gateways? Past a few, a fleet layer like EAM pays off.
- Alerts or answers? Alerting needs thresholds and pipelines. Root cause needs deep per gateway diagnostics.
- Which Ignition version? Catalyst needs 2027, Gateway Analyzer needs 8.3.2 or newer, EAM and System tags work on both 8.1 and 8.3.
- AI and data policy: if gateway data cannot leave the site, plan a self hosted model before relying on Catalyst.
- Who reads the output? Operators need a screen, engineers need detail, and customers or support teams need a report.
Figure 3. A simple way to choose. Most multi gateway teams end up combining a fleet layer for alerts with a per gateway layer for root cause.
When a dedicated gateway analyzer tool helps
Built in tools show you a lot, but the signals are spread across many pages. The Plantformance Gateway Analyzer module is one option for pulling them together. Based on its product page and documentation, it:
- Groups databases, tags, scripts and network resources into a single Gateway Health Score with an actionable findings report.
- Covers resources, datasources, scripts, sessions, tags, security, alarms, audit, projects, redundancy, Gateway Network and store and forward.
- Shows findings grouped by severity on a Summary page under Diagnostics > Gateway Analyzer, with an Analysis page to drill into each area.
- Exports JSON and CSV reports, and can schedule report delivery to HTTP, web socket, file transfer and email destinations.
- Creates an optional managed diagnostic tag provider so projects can display or react to analysis values.
Be clear about what it is not. Its documentation states that it analyzes the local gateway, is not a centralized multi gateway management product, and does not guarantee every issue is detected. For a fleet, use EAM or System tag alarms as the central alerting layer and install the analyzer where you want deeper reviews, with scheduled reports going to one shared destination.
It is licensed per gateway, and anyone can run it in a 2 hour trial mode that can be reset at no cost. The trial shows the overall health score and up to 5 summarized findings, hides sensitive diagnostics and blocks configuration changes and scheduled delivery. The full version adds risk scoring, trend analysis, forecasting and dependency mapping.
A central gateway cannot read Gateway Analyzer data from other gateways: the module has no Gateway Network support and analyzes only the gateway it is installed on. Gateway Analyzer has no scripting functions: you work with its results on its Gateway pages and through its reports and exports.
Plantformance is an Inductive Automation Alliance Partner that works only on Ignition; see the Ignition modules overview for the other modules. If you would rather not run monitoring yourself, the managed Ignition on premises service includes fleet management and a 24/7/365 Operations Center. We also cover the Ignition gateway security audit checklist, managed Ignition services for multi-plant teams and how to choose third-party Ignition modules. To plan a monitoring setup for your gateways, talk to the team.
Frequently asked questions
What is the best tool to monitor multiple Ignition gateways?
For alerts across many gateways, the Enterprise Administration Module is the built in choice: one controller gateway watches every agent, stores Agent Events and raises alarms on CPU, memory, errors and database use. Pair it with a per gateway tool, such as the Diagnostics pages, Catalyst or an analyzer module, to find root causes.
Is Catalyst in Ignition 2027 enough for gateway diagnostics?
Catalyst Core adds Diagnostic Assistants that summarize issues on gateway pages such as Threads, Database Connections and Store and Forward, and suggest potential fixes. It needs a configured AI model provider and works page by page. It is not a fleet monitoring or alerting system, so most teams still need EAM or System tag alarms.
How do I get an alert when an Ignition gateway loses its database connection?
Add an alarm to the System tag [System]Gateway/Database/<connection name>/Available that triggers when the value is false, then route it to an alarm notification pipeline with an on call roster. Add a second alarm on ConnectionSaturation for early warning. With EAM, enable Metrics Alarms to watch database pool utilization on every agent.
What memory threshold should I alarm on for an Ignition gateway?
Inductive Automation's only published numbers are the EAM defaults: a warning at 70% and an error at 90% of the agent's memory use. They are starting points, not rules. Watch Memory Utilization and Memory Usage against Max Memory for a few normal weeks, then set alarms that fit that gateway's baseline.
Can Plantformance Gateway Analyzer monitor several gateways from one screen?
No. Its documentation says it analyzes the local gateway and is not a centralized multi gateway management product. Install it on each gateway you want reviewed, schedule its reports to one shared destination, and use EAM or System tag alarms as the central alerting layer. It requires Ignition 8.3.2 or newer.
Talk to a Plantformance Ignition engineer
Tell us about your Ignition setup and we will point you to the right module or managed service for your plants.