Every connector has a page in the directory with the same six parts. This page explains those parts, lists the prerequisites administrators most often need to grant, and describes the connection registry and the vault.
| Part | What it tells you |
|---|---|
| Methods | Each way Lima reads or writes the system, what it does and when it is used. |
| Objects | The tables, views or records reached, with their plain meaning. |
| Patterns | Full load, incremental, change capture, streaming, write-back: which are supported and how. |
| Prerequisites | What the team that runs the system grants: users, authorisations, network, caps. |
| Compliance | For SAP connectors, the released interfaces used and the restricted ones avoided. |
| Pair pages | The source-to-target paths this connector appears in. |
The connector directory: 60 connectors and 150 pair pages →
| Category | Typical prerequisites |
|---|---|
| SAP | A communication user with read authorisation on the tables, function modules and BAPIs in scope; PrivateLink, VPN or a jump host; a session cap and run windows agreed with Basis; for write-back, BAPI authorisations and a test client. |
| Oracle | A read-only database user on the schemas in scope; a connection cap and run windows agreed with the DBA; supplemental logging for LogMiner; for write-back, a responsibility for the import programs. |
| Databases | A user with SELECT (plus replication privileges for change capture); the change mechanism enabled by the DBA; retention for the run window. |
| Warehouses and lakehouses | A role or service principal with create rights on the target schema; a stage or landing bucket; keys. |
| Files | Host, credentials or key, folder permissions, name patterns, the archive rule. |
| Streams | Bootstrap servers or endpoints, credentials, topic or subscription permissions, schema registry access. |
| Applications | An integration user or connected app with the object permissions; API limits reviewed. |
Credentials are stored once in the vault and referenced by name from connections. A pipeline definition never contains a key, and a log never prints one. A connection is tested when it is registered and every time its credentials rotate; a failed test names the cause (authentication, network, permission) so the right administrator is asked.
