Connectors
Connectors
A connector is a tested connection to your database. Here is what InfiniStack connects to today, how a connection is added and proven, and the role each one plays.
The credentials are the easy part
Storing a host and a password takes five minutes. The work is everything around it: proving the connection actually opens, using only the database types your organization is allowed to connect, deciding whether it is read from, written to, or both, and keeping every organization’s connections apart. A connector here covers all of that—not just the login.
- Before it saves
- Test the connection
- Types
- Allowed per org
- Role
- Source, destination, or both
- Availability
- Scoped per org
What you can connect today
InfiniStack connects to two relational engines, and does both properly: PostgreSQL and MySQL (including MariaDB). Each works end to end as a source, a destination, or both on the same connection—from the connection test onward.
PostgreSQL
Connect to a PostgreSQL database. Schema-aware, so tables in sales and public stay distinct rather than colliding on name. Choose which schemas to read; public is the default. Port defaults to 5432.
MySQL & MariaDB
Connect to a MySQL or MariaDB database. The same form and the same connection test as PostgreSQL. Port defaults to 3306.
Setting one up
Two steps
Adding a connection is a short walk, and each step ends with something confirmed rather than assumed—a type your organization is allowed to use, and a connection that actually opened.
Pick a type from what your organization is allowed
The picker shows, before you click, whether each type can act as a source, a destination, or both. You only ever see types a platform admin has enabled for your organization—there is no route to connect something outside that list.
Fill in the form — then prove it works
Host, port, database, user, password, SSL mode. The port is pre-filled from the type you picked, so 5432 and 3306 are one less thing to look up. Test connection is right there on the form, and the result is stored with the connector.
Under the connection
What happens when you save one
A connector is not a stored login that a shared service borrows. Saving one sets up what is needed to reach your database, inside your organization’s own environment.
Open connectors
Reads and writes go through open, standard connectors—the same approach used across the data-tooling world, not a proprietary driver.
Isolated per organization
Each organization gets its own isolated environment. What a connector needs is set up there in the background, and one organization’s connections are never visible to another.
Tested against the real database
Test connection opens a real session to your database, so you find out about a wrong host, port, or password before the connector is saved.
Source, destination, or the same one twice
Every connection declares a role. It is not cosmetic—it decides where the connector can be used, so picking the wrong end is something you find out at setup rather than at runtime.
Source
Data is read out of it. This is the database your application writes to, and the one you would rather not have analysts querying directly.
Where your data comes from
Available as a data source
Readable by workflows and chat
Destination
Data is written into it. This is the analytical side—the place dashboards, metrics, and transformations are allowed to read from without touching production.
Available as a data destination
Where transformations land
What dashboards query
Both
The same connection on either end. Useful when one database both receives raw loads and feeds the modelled tables built on top of them.
One set of credentials
Read and write in one place
Both roles, one connection test
Common questions
Why only PostgreSQL and MySQL?
Because a connector is only listed once it works the whole way through—connection test, schema awareness, and use as a source or a destination. Two databases that do all of that are more useful than fifty that stop at the login screen. PostgreSQL and MySQL are what InfiniStack connects to today.
Who can see my connection details?
Enabling a connector type and supplying credentials for it are two different jobs. A platform admin enables the type for your organization; the connection itself belongs to your organization’s own record, inside your organization’s own project. Nothing about a connector is shared across tenants.
Can I connect the same database more than once?
Yes. A connector is a named connection, not a database—so a read-only replica for one purpose and a separate, narrower login for another can live side by side, each with its own credentials.
Can one connection be both a source and a destination?
Yes. Both supported engines can be set up as a source, a destination, or both on the same connection—one set of credentials, one connection test.
What if my connection test fails?
You see it on the form, before anything is saved. Check the host, port, database name, credentials, and SSL mode, then run the test again.
Do I need a data warehouse before I can start?
No. Both supported engines work as destinations as well as sources, so a second database is a perfectly good place to land your data.
Connect it once, use it everywhere
One tested connection feeds the pipelines, the workflows, and every dashboard built on top of them.