Glossary
KRISID's own words, defined once.
KRISID uses some everyday words in a particular way. Each is defined here once, with a link to the page that covers it. Your repository may use its own names for types, attributes and roles; those come from its metamodel.
A to D
Asset. Anything KRISID records: a business term, a domain, a table, a column, a report, a metric. Each asset has a type and a page of its own. See What KRISID is.
Asset type. The kind of thing an asset is, such as Business Term or Table. The metamodel decides which types exist. See The metamodel.
Attribute. A fact recorded on an asset, such as its Definition, its Status or whether it Contains PII. See Reading an asset.
Authored. Written and maintained by people in KRISID, as opposed to harvested. Business terms and domains are usually authored. See What KRISID is.
Approver. Someone who may take a step out of a waiting state, usually because they hold a role over the asset. A state can ask any one of them or every one of them. See Requests and approvals.
Best match. The one result Ask shows first, in its own card, when it is plainly the answer: its name is what you typed, it is the only result, or it is well ahead of the next. See Ask: search and questions.
Catalogued by. The source that first brought an asset in. It is recorded once and does not change when a second source later reports the same object, so counting what a source catalogued gives the same answer tomorrow. See Connections and harvesting.
Change set. Your edits to one governed asset, collected on its page and submitted together as one request. It's kept as a draft until you submit or discard it. See Requests and approvals.
Condition. A rule that narrows a list, such as Status is Approved. You add one with + add a condition under an answer. The same conditions decide a workflow's scope and the paths out of a condition box. See Lists, conditions and saved searches.
Condition box. A state in a workflow where nobody decides. The request goes on by itself, along the first path whose conditions hold, or along the otherwise path. It is drawn as a diamond. See Requests and approvals.
Connection. A source system KRISID reads from, such as a Snowflake account or a Power BI workspace, with the details it needs to reach it. Administrators manage connections on Sources. See Connections and harvesting.
Context layer. What KRISID keeps about each asset so search can find it: its own words and the words of what surrounds it, such as the term it realises or the domain it belongs to. See What KRISID is.
Demonstration data. Invented example assets that may come with a new repository. They are labelled demo data and always listed after real results. See Troubleshooting and FAQ.
Delete harvested assets. The permission (HARVESTED_DELETE) that removing an asset a harvest brought in needs, on top of Delete assets. Only Sysadmin has it in the starter model. See Editing an asset.
Dimension. Something a metric can be sliced by, such as region or channel. It is part of the semantic model. See Metrics and asking for a number.
Domain. An area of the business, such as Customer or Finance, that answers for a set of assets. In the starter model it is the Data Domain type. KRISID also calls this a grouping. See Browsing by domain and by type.
G to M
Estate. A layer of the repository browsed as it nests. Browse the physical estate goes from source systems down to columns. Browse the logical estate goes from Subject Areas to Data Entities to Data Attributes. The top level of each can be filtered as you type. See Browsing by domain and by type.
Grouped by. How far out a layer of the diagram is grouped: every asset drawn as the thing that holds it at the level you pick — Table or View, Schema, Database, Source System. The steps are whatever your metamodel nests, one choice per layer, and nothing is lost. See Lineage and impact.
Grouping. A type whose assets gather others under them by ownership, so that you can browse By it. Data Domain is the usual one. See Browsing by domain and by type.
Harvest. Reading a source system through a connection and recording what is in it as assets. You can preview a harvest before it writes anything. See Connections and harvesting.
Harvested. Created by a harvest rather than typed in. A harvested asset's name and key come from the source, so you cannot change them in KRISID. See What KRISID is.
Intake key. The credential an OpenLineage source receives events with, instead of a password. It is shown once when it is made, then only as dots with its date. Making a new one stops the old one working. See Connections and harvesting.
Harvested attribute. An attribute written by a harvest from what the source system says, such as Owner at Source or Description from Source. It filters and searches like any attribute but cannot be edited in KRISID. See Reading an asset.
Hierarchy view. A saved view of one type with levels under it, each reached through a relation you choose and with its own columns. It can be exported to Excel as an outline, one row per asset under its parent, grouped so each branch folds. See Browsing by domain and by type.
Lineage. The path data takes from one asset to another, drawn as a picture you can follow upstream or downstream. See Lineage and connections.
Mark. A sensitive value shown as a label on an asset wherever it appears, such as Contains PII. Which values count as sensitive is declared in the metamodel. See Classifications and sensitive data.
Metamodel. The description of what your repository may hold: its asset types, attributes and relations. It is why your screens use your organisation's words. See The metamodel.
Metric. A defined business number, such as Net Revenue, in the semantic model. Ask can compose a query for it but never runs one. See Metrics and asking for a number.
Missing in Source. A source status. It means the last harvest that looked did not find this asset. Nothing has been deleted, and the status returns to In Source as soon as the source shows the asset again. Marked assets can later be cleared in bulk, once they have been missing for longer than the cut-off. See Troubleshooting and FAQ and Clearing up what has gone.
O to R
Order. Which workflow is tried first when several start on the same type and event. Lower is first. See Requests and approvals.
Owner. Short for Data Owner in the starter model: the role accountable for an asset, who decides who else is responsible for it. See Owners, stewards and responsibilities.
Persona. Another word for role, as in the People and Persona screen. See People, roles and access.
Proposed asset. A new asset of a governed type, waiting for approval. Only its proposer, its approvers and administrators can see it. See Requests and approvals.
Relation. A link between two assets, such as owned by. It shows on both assets' pages. See What KRISID is.
Request. A proposed change waiting for somebody's decision, because a workflow governs that kind of change. Requests are listed under Change requests. See Requests and approvals.
Responsibility. A record of who answers for an asset and in what role. A person or a team can hold one, and it can be inherited from a domain or container. See Owners, stewards and responsibilities.
Role. What somebody may do, either everywhere (such as Viewer) or over one asset (such as Data Steward). See People, roles and access.
Run. One thing a source did: a preview, a sync, or the catalogue or lineage pass inside a sync. Each keeps when it started and when it ended, and each has a page of its own listing every line it wrote — including a run that wrote none. See Connections and harvesting.
Roll-up. Everything a domain answers for: what it owns directly, plus everything inside those things. A column rolls up to its table's domain. See Browsing by domain and by type.
S to W
Reminder. A message sent to the approvers who haven't answered yet, by email and on their Change requests screen, every so often or a set time before the limit. See Requests and approvals.
Saved search. A search or list you kept with Save this search, to open again or share. See Lists, conditions and saved searches.
Scope. The assets of its types a workflow is for, set under Only for these as groups of conditions. For example, Data Entities that are not in the Finance or Customer domains. See Requests and approvals.
Secret. A password, private key, token or client secret a source signs in with. It is stored encrypted and shown only as dots, with when it last changed and who changed it. Replace it with Change sign-in details when it is rotated. See Connections and harvesting.
Segment. A named, reusable filter in the semantic model, such as completed orders only or external customers only. See Metrics and asking for a number.
Semantic model. The metrics, dimensions and segments your organisation has defined, and how they map to tables. Ask builds queries from it. See Editing the semantic model.
Source connection. An established source on Sources: how KRISID signs in to a system and what it brings. Everybody sees the list; opening one needs Open source connections (SOURCE_VIEW). What it brought in is readable by all. See Connections and harvesting.
Signed in from. Where a sign-in came from: the network, masked (192.168.4.0/24), and the name this repository gives it. Off until a Sysadmin switches it on, and named from a list kept here — nothing is looked up anywhere. See Usage analytics.
Maintenance page. The single page everybody sees while the repository is closed for work. Switched on People and Persona → Maintenance by whoever may manage access; they keep working, everybody else is signed out. See People, roles and access.
Source system. The asset standing for a connected system in the physical estate. It is called whatever you named the source, and its key is what the connector reported. See Connections and harvesting.
Today (in a condition). A date condition can be compared with today, today + 30 days or today - 7 days rather than a fixed date. It is worked out when the list is opened or the request is raised. See Browsing by domain and by type.
Steward. Short for Data Steward in the starter model: the role that looks after an asset's meaning and quality day to day. See Owners, stewards and responsibilities.
Sync. A harvest that writes. Sync now records what the source holds; a preview only shows what a sync would do. See Connections and harvesting.
Time zone. The zone a person's browser last reported, kept on them and shown read-only on People and Persona. Deadlines are written in it — on screen and in that person's own email — always labelled, so one moment reads 21:30 in Mumbai and 12:00 in New York. Nothing counted changes with it. See Change requests.
Time allowed. How long a waiting state gives its approvers, in hours or days, optionally counting working days only. When it runs out, the workflow takes the arrow chosen for it. See Requests and approvals.
Tag push. Writing classifications from KRISID onto Snowflake as tags. It is the only thing KRISID writes to a source, and it is previewed first. See Pushing tags back to Snowflake.
Type group. A band of asset types, such as Business or Physical, as the sidebar groups them. See Finding your way around.
Usage analytics. A record of who opened which screens, what they searched for and what they changed. It is kept only while a Sysadmin has it switched on. See Usage analytics.
View as. Looking at the repository as somebody else sees it, to check what they can reach. Nothing can be changed while you do. See People, roles and access.
Workflow controls. Whether a direct API write or a workbook load on a governed asset is written, proposed or refused. They are set on People and Persona. See People, roles and access.
Workflow. The rule that says which changes need a decision, who decides, and in what order. It has a name you can change, and a code and UID that never change. See Requests and approvals.