The Deployment View
Where Infohub's running components are defined — platforms, servers, engines, connections and consumers — and the rules about what can be added where.
The Model, Template and Report views describe what Infohub calculates and shows. The Deployment view describes what runs it: which hosts exist, which engines live on them, and which connections and consumers each engine owns.
If the model is the recipe, this is the kitchen. A model can be entirely correct and produce nothing at all because no engine was ever told to own it.
The shape of the tree#
Deployment objects nest in a fixed hierarchy:
Platform a Flow host
+-- Server the web client and REST API
+-- Data Engine retrieval and calculation
| +-- Connection a configured data source
+-- Integration Engine publishing to consumers
| +-- Consumer a configured destination
+-- Message Engine limits, messages and notifications
+-- Message Channel
+-- Notification Server
Template Server a federated template source (top level)
Platforms and Template Servers sit at the root. Everything else hangs off a platform. You cannot add a Data Engine without first having a platform to put it on.
A remote platform cannot host engines.
Servers can be added to any platform. Data Engines, Integration Engines and Message Engines can only be added to a local platform — the commands are disabled on a platform marked remote.
A remote platform represents an Infohub host that another installation owns. You add it so this instance can reach its Server; you do not run engines on someone else's host. In a multi-site deployment this is the constraint that decides where processing lives — see Multi-site Architecture.
Adding components requires administrator#
Every Add command in this view — platform, server, any engine, template server — is available only to a user with administrator rights. A non-administrator sees the tree and its state but cannot extend it. If the Add commands are missing entirely rather than greyed out, check the signed-in user's rights before looking for anything else, as covered in Users, Groups and Access.
What each component does#
| Component | Responsibility |
|---|---|
| Platform | The host process. Owns the engines and servers that run on one machine, and coordinates them |
| Server | Serves the web client and the REST API — see Flow Server API Reference |
| Data Engine | Retrieves from sources and runs calculations for the metrics assigned to it |
| Integration Engine | Publishes values out to consumers |
| Message Engine | Evaluates limits and sends messages through notification services |
| Template Server | A federated source of templates — see Template Servers and Federation |
| Connection | A configured data source instance, owned by a Data Engine |
| Consumer | A configured destination, owned by an Integration Engine |
For what these components are architecturally, rather than how they are configured, see Infohub Architecture.
Branding the Flow Server#
Double-click the Server node and its properties open, including the ones that decide how the web client looks:
| Property | Sets |
|---|---|
| Name and Description | How this server is identified |
| Themes | The theme applied to the web client — colours together with images |
| Colors | The theme's colour set, which charts can reference |
The colours are worth understanding as more than decoration. Because charts can reference the theme's colours rather than hard-coding their own, setting them once here gives every chart on the server a consistent palette — and a palette that matches corporate branding, which is usually what is being asked for when somebody wants Infohub "in our colours".
Doing it the other way round — picking colours per chart — produces a server where forty dashboards nearly agree, and where a brand refresh means editing all forty. See Colour as Information for what colour should be carrying once it is consistent.
Defaults#
Four deployment types can be marked Set as Default: Data Engine, Integration Engine, Message Engine and Platform. The default is what new model content is assigned to when nothing is chosen explicitly.
Servers and Template Servers have no default — there is nothing that gets silently assigned to them.
As in the Toolbox, the command is only enabled on an object that is not already the default. Setting the default Data Engine early matters for the same reason the default time scheme does: everything created afterwards inherits it, and correcting the assignment later is a per-object job.
Deployment state#
Every deployment object carries a state, and Change deployment state is how you deploy or undeploy it. The states and the workflow are the same as for model content — see Deploying Configuration, which covers the state table, the order that works, and the Platform host check.
Two operations specific to this view:
- Configure connection — opens the connection properties for a Connection object, the same property grid described in Connection Property Reference.
- Migrate template server — moves a template server registration, used when a federated template source changes host.
When the host refuses a deployment#
Deploying a Platform registers the instance with the Flow Bootstrap service on that host, which checks that it can reach the instance database before it accepts. When Bootstrap refuses, Flow Config now reports the refusal with Bootstrap's own reason rather than reporting success and leaving the platform undeployed, or showing the older "please install it manually" message with no cause.
The most common refusal is a login failure. An instance on a SQL Server login that was deployed with Windows authentication — the default until a Username and Password are entered — makes Bootstrap's services connect as the service account, which SQL Server rejects. The message says so and points at the fix: open System properties and enter the SQL login as Username and Password, then deploy again. The same check runs when a script registers with --register-with-bootstrap, which exits non-zero with the same text.
The view no longer waits on an unreachable platform. Each platform's version and running components come from a single, time-bounded probe of its host. A host that is off the network shows as not responding; it no longer stalls Flow Config's startup or blanks the components of the platforms that are reachable.
Monitoring what is running#
Monitor activity opens live activity for the selected object, and Task opens a running Server in a browser or copies its link. Task is enabled only while the parent Server is actually running, which makes it a quick liveness check in itself.
For health beyond the tree, see Monitoring Infohub Health and Reading Infohub Logs.
System-wide settings#
Instance-level settings live in System properties, not on any single deployment object: the instance name, the bootstrap port of the host, whether daily database backups run, and the hour database maintenance starts.
Those last two are consequential and are covered in full elsewhere — see The Maintenance Window and Backup and Recovery.
Diagnosing#
| Symptom | Likely cause |
|---|---|
| Part of the model produces nothing | It is assigned to no Data Engine, or to one that is not deployed |
| Add Data Engine is greyed out on a platform | The platform is remote; engines run only on local platforms |
| All Add commands are missing | The signed-in user is not an administrator |
| Task is greyed out on a Server | The Server is not running |
| A new measure went to the wrong engine | The default Data Engine is not the one you expected |
| Edits are not taking effect | The object's state is Different; deploy it — see Deploying Configuration |
| A template server stopped resolving after a host move | The registration needs Migrate template server |
| Deploy reports "Flow Bootstrap on … refused the registration" with a login failure | The instance uses a SQL login but none is entered under System properties; enter Username and Password and deploy again |
| A platform shows no version and no running components | Its host did not answer the probe; check that Flow Bootstrap is running there and the port is reachable |