You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
We're excited to announce the release of CHouse UI v3.7.0!
What's New
Added
Runtime Lineage (Scheduled Queries) — a new Runtime Lineage sub-tab that shows observed-runtime data lineage for a selected job. Reads system.query_log (every run is tagged with its job_id) to graph the tables a job actually reads and writes — chaining jobs together when one job's destination table is another job's source — and reveals the columns observed flowing through each table/job when a node is selected (column level). The graph expands on demand — each card shows one level upstream/downstream and a + reveals the next hop in that direction — and uses the same searchable job filter as Runs.
Scheduled Queries (DataOps) — a new top-level DataOps page with a Scheduled Queries feature (Overview / Jobs / Runs). Schedule any read-only SELECT on a daily/weekly/monthly preset or a custom UTC cron expression, with deterministic time windows ({{slot_start}}/{{slot_end}}/{{prev_run_at}}), bounded result snapshots, and an optional engine-generated, idempotent materialize write-back (append / replace-partition / upsert) into a destination table. Failure-based alerting notifies the linked notification channels when a run fails (and once on recovery), transition-based to avoid flapping. The Overview is a real summary of the feature (health KPIs, success rate, cadence / output-mode / last-run breakdowns, upcoming runs, top failing jobs); the Jobs list is filterable (enabled/disabled, last-run state, name search) and the Jobs and Runs lists are paginated. Includes an in-process per-job-lease scheduler that is correct under multiple replicas with no leader election, a crash-only reaper + bounded retry, and a transactional notification outbox for at-least-once delivery. Gated by new scheduled_queries:view|edit|delete|run|write RBAC permissions. Opt-out via SCHEDULED_QUERIES_ENABLED=false.
Fixed
SSO no longer silently escalates privileges ([Bug]: sso auto maps role with highest priviledge #270) — when an IdP claim resolved to more than one mapped role, the previous behaviour collapsed to the highest-privilege match, so a user in multiple groups could land in an unexpectedly powerful role. Role sync now fails closed: an ambiguous claim assigns no role and keeps the user's existing one, logging a warning so the misconfiguration can be fixed. Multi-group role mappings remain supported — only genuine overlap (one user resolving to several roles) is rejected.
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
Uh oh!
There was an error while loading. Please reload this page.
🎉 CHouse UI v3.7.0 Released!
We're excited to announce the release of CHouse UI v3.7.0!
What's New
Added
system.query_log(every run is tagged with itsjob_id) to graph the tables a job actually reads and writes — chaining jobs together when one job's destination table is another job's source — and reveals the columns observed flowing through each table/job when a node is selected (column level). The graph expands on demand — each card shows one level upstream/downstream and a+reveals the next hop in that direction — and uses the same searchable job filter as Runs.SELECTon a daily/weekly/monthly preset or a custom UTC cron expression, with deterministic time windows ({{slot_start}}/{{slot_end}}/{{prev_run_at}}), bounded result snapshots, and an optional engine-generated, idempotent materialize write-back (append / replace-partition / upsert) into a destination table. Failure-based alerting notifies the linked notification channels when a run fails (and once on recovery), transition-based to avoid flapping. The Overview is a real summary of the feature (health KPIs, success rate, cadence / output-mode / last-run breakdowns, upcoming runs, top failing jobs); the Jobs list is filterable (enabled/disabled, last-run state, name search) and the Jobs and Runs lists are paginated. Includes an in-process per-job-lease scheduler that is correct under multiple replicas with no leader election, a crash-only reaper + bounded retry, and a transactional notification outbox for at-least-once delivery. Gated by newscheduled_queries:view|edit|delete|run|writeRBAC permissions. Opt-out viaSCHEDULED_QUERIES_ENABLED=false.Fixed
📦 Installation
🔗 Resources
All reactions