@overkube/query

What overkube asks of a cluster's data, as data — pure, with no dependencies: the page answers from what it watches, a server could from its indexes, and nothing that asks has to know which.

• fields (FieldDef, defineField, checkFieldDef, fieldsOf): a value by a path into the object, worked out of it by CEL, or joined (Join: a field of another kind's object of the same namespace and name — a pod's metrics; or, keyed (by, to, reduce), of those of its namespace whose field to is its own by, reduced to the least (a time: the soonest), the greatest (of strings, none of them numbers: the least or greatest), a sum or how many — a workload's reports, an Ingress's certificates' soonest end); its type, unit, how it sorts;
• queries (GetQuery, ListQuery, CountQuery, SearchQuery, LogQuery): one object, a window of a list (Where, Sort, Page), a count by a field, a search, a log's window; checked (checkQuery) and bounded (LIMITS, FIELD_LIMITS);
• results (Result, Live, State): live while read, with their state and errors;
• access (AccessRule, allowed, ruleAllows): what one may do, as a SelfSubjectRulesReview tells it;
• a scan (scan, scanCount, windowOf, matches, ordered): a query answered over the objects given; a log's window picked out of its lines (logWindow, entriesOf);
• kubectl's JSONPath (compilePath, valuesAt, isPath), and canonical hashing (canonical, fnv1a).

Used by the source (the answers), the connector, the gateway, cel and kube-fields. npm run stress -w @overkube/query runs the scan's stress test.

Built with npm run build -w @overkube/query (dist/); in the workspace read from src/.