* refactor: migrate from pg-transact to effectful-postgresql
Replace pg-transact and pg-transact-effectful with effectful-postgresql for database operations. Key changes: replace DB effect with WithConnection, remove dbtToEff wrappers, use direct PG.query/PG.execute calls, update runDB to runWithConnectionPool, keep pg-entity for query building, add IOE constraint alongside WithConnection where needed.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
* Auto-format code with fourmolu
* chore: use effectful-postgresql fork with OpenTelemetry support
Switch to tonyalaribe/effectful-postgresql fork (instrument-pg branch) and
enable the +enable-otel flag for OpenTelemetry instrumentation.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
* fix: improve connection usage and error handling in BackgroundJobs
- Use PG.execute directly instead of multiple withConnection calls for
CleanupDemoProject (more efficient connection usage)
- Handle missing project gracefully in ATRuntimeException processing
instead of crashing with error
- Clean up unused imports (PGS, withConnection)
🤖 Generated with [Claude Code](https://claude.com/claude-code)
* fix: correct effectful-postgresql subdir path
🤖 Generated with [Claude Code](https://claude.com/claude-code)
* fix: add proper error handling and logging for critical operations
- Log advisory lock release failures instead of silently ignoring
- Log when error update doesn't match any span (potential data issue)
- Log when span hash update count mismatches expected (data integrity)
🤖 Generated with [Claude Code](https://claude.com/claude-code)
* Auto-format code with fourmolu
* refactor: simplify single-row query handling in BackgroundJobs
* refactor: return lists from model query functions instead of Vectors
PostgreSQL-simple returns lists naturally, so wrapping results with V.fromList was wasteful. Model query functions now return lists directly, with Vector conversion happening at call sites that actually need Vectors.
* Auto-format code with fourmolu
* perf: fix N+1 query pattern and improve advisory lock handling
- Replace individual UPDATE queries in forM_ loop with batched UPDATE
using SQL unnest to reduce O(n) DB round-trips to O(1)
- Add 5-minute statement timeout to advisory lock to prevent hangs
- Add timeout reset in finally block for cleanup
- Improve comments explaining session-level lock auto-release behavior
* refactor: replace verbose Only unwrapping with coerce
- Use coerce @[Only T] @[T] instead of map (\(Only x) -> x)
- Use coerce @(Maybe (Only T)) @(Maybe T) instead of fmap (\(Only x) -> x)
- Update GenerateOtelFacetsBatch to use Projects.ProjectId instead of Text
* Auto-format code with fourmolu
* fix: advisory lock release and remove unnecessary vector conversion
- Use query instead of execute for pg_advisory_unlock (SELECT returns result)
- Add validation for unexpected unlock results
- Remove unnecessary V.fromList/V.toList round-trip in runHourlyJob
* refactor: simplify COUNT(*) query result handling
Use inline maybe/listToMaybe instead of case expression for cleaner code.
* Auto-format code with fourmolu
* refactor: use catch instead of try/case for advisory lock cleanup
Cleaner error handling with catch pattern instead of explicit try + case Left/Right.
* Auto-format code with fourmolu
* perf: use V.mapMaybe and V.unzip3 to avoid list conversions
Direct Vector operations for batch error updates instead of
list mapMaybe + V.fromList.
* refactor: introduce DB type alias and switch to pg-transact
* Auto-format code with fourmolu
* fix errors in specs
* fix: remove duplicate import in Dashboards.hs
* fix: resolve integration test failures
- BusinessFlowsSpec: fix column names for apis.subscriptions table
- TestUtils: add test user as admin to nil UUID project
- System/Types: increase static UUID pool to prevent collisions
- AnomaliesSpec: fix test assertions for correct filter views
* fix aroundAll
* fix: resolve integration test failures
- Fix insertProjectMembers to handle empty list (prevents 'no rows supplied' error)
- Fix DashboardsSpec UUID collisions by grouping creations and adding cleanup
- Fix ManageMembersSpec: correct userID to UUID 1, update expected member count and email
- Fix ProjectsSpec: update expected project count to include demo project
* remove PG_EFFECT_MIGRATION.md
* fix: use unique team handles in bulk add teams test to avoid conflicts
---------
Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
* Improve dashboards and teams UI with star functionality
- Add star/unstar functionality for dashboards with persistent state
- Update dashboard sorting to prioritize starred items
- Improve table component styling and layout
- Add icons for dashboards and teams in list views
- Streamline table columns and remove unnecessary checkboxes
- Update bulk actions to use new table component features
* Auto-format code with fourmolu
* Fix teams bulk actions to use itemId field
Update TBulkActionForm to use itemId instead of teamId to match
the Table component's standardized field name for bulk actions.
* Improve settings navigation layout and styling
- Reduce navigation width for better space utilization
- Update font sizes and spacing for cleaner appearance
- Improve icon styling with background colors
- Enhance navigation item padding and layout
* Refactor API keys page to use Table component and add client-side tab support
- Enhanced TabFilter component to support both client-side and server-side tabs
- Added clientSide boolean field to TabFilter for tab switching mode
- Added targetId field to TabFilterOpt for custom element targeting
- Refactored API keys page to use reusable Table component
- Updated all existing TabFilter usage to include new fields
- Improved consistency across table implementations
* Auto-format code with fourmolu
* nitpicks on
* add test for starring and unstarring, and dashboard default sorting
* Fix tagify scope issue in dashboard creation modal
The tagify variable was scoped inside DOMContentLoaded event listener,
causing a reference error when getSelectedTeams() tried to access it
during form submission. Moved declaration to outer scope.
* Fix gridStackInstance null error when adding dashboard widgets
* Fix teams badge WCAG contrast by using badge-neutral
* Fix compilation errors
- Fix do-notation errors by adding proper do blocks instead of using >> after $
- Rename local section_ helper to teamSection_ to avoid shadowing
- Remove unused ProjectId parameter from renderMonitorContent
* Use brackets instead of do-blocks for single-line expressions
- Keep code on one line by using parentheses and >> operator
- Applied to card_, notifRow_, lazySection_, teamSection_, field_, fieldIcon_ helpers
* Apply code formatting updates
- Replace font-mono with monospace class
- Minor whitespace adjustments
* Fix duplicate html assignment bug and simplify code
* Auto-format code with fourmolu
* Refactor tables to use standardized Table component with sorting
- Add containerId to Table Config for HTMX targeting
- Add renderAsTable mode for proper table rendering
- Add rowId and rowAttrs to TableRows for load-more consistency
- Fix th elements to include column attrs for alignment
- Fix class concatenation spacing in table headers
- Fix scrolling in BodyWrapper (overflow-y-hidden to overflow-y-auto)
- Add sort parameter to Issues.selectIssues
- Refactor Anomalies and ApiCatalog to use TableHeaderActions
- Remove flex classes from table cells
* Auto-format code with fourmolu
* Fix table load more by moving pagination inside tbody
- Move pagination link inside tbody for table mode (as a <tr>)
- Add renderTablePaginationLink with hx-target='closest tr'
- Add id to tbody for targeting
- Pagination row gets replaced with new rows + new pagination row
* Refactor Anomalies table to use built-in checkbox
- Remove custom checkbox column (use table's rowId feature instead)
- Add status indicator dot to Issue column title row
- Update TableRows to use proper rowId for load more
* Auto-format code with fourmolu
* Fix Anomalies issue column layout structure
- Fix indentation so metadata, stats, actions are separate blocks
- Status indicator dot in title row
- Condense code using single-line patterns where appropriate
* Auto-format code with fourmolu
* Merge renderPaginationLink and renderTablePaginationLink into single function
* Auto-format code with fourmolu
* Align row checkboxes to top and add vertical padding to issue column
* Improve settings sidebar: larger text, neutral icons, stronger selection highlight
* Widen settings sidebar
* Fix hlint warnings
* Fix all hlint warnings across codebase
* Auto-format code with fourmolu
* Fix GHC warnings: dodgy import, name shadowing, unused do-bind
* Use foldMap and zipWithM_ for cleaner iteration patterns
* Extract columnCount helper and add error handling for processVariable
* Auto-format code with fourmolu
* Optimize URL construction and use single-pass fold for change counting
* Extract issueTypeBadge helper to reduce code duplication
* Fix SQL injection vulnerabilities and semantic error in change counting
* Auto-format code with fourmolu
* Simplify SQL validation and remove redundant variable
* Auto-format code with fourmolu
* Rename selectDashboardsSorted to selectDashboardsSortedBy for clarity
---------
Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
Update all integration tests to properly unpack tuple response from testServant helper, which now returns (RespHeaders a, a).
Key changes:
- Add servant and servant-htmx to integration test dependencies
- Fix tuple unpacking in all test specs (AnomalyList, Log, Alerts, ManageMembers, Projects, BusinessFlows, Reports, Api, ApiCatalog, GrpcIngestion)
- Fix BusinessFlowsSpec imports to use ResponseHeader data constructor
- Fix type-level string usage for header lookups
- Fix ProjectId to Text conversions using .toText
- Add missing pattern match cases to avoid incomplete pattern warnings
- Add openai package dependency
- Update LLM effect to use OpenAI V1 chat completion parameters
- Pass model name and messages via proper parameter structure
- Fix error handling to use show instead of toText
- Add GrpcIngestionSpec to test suite in package.yaml and cabal
- Add required dependencies: lens, proto-lens, grapesy
- Update Makefile timescaledb-docker-tmp to use faketime for frozen test time
- Replace O(n²) nubBy with O(n) HashMap-based deduplication for endpoints, shapes, fields, and formats
- Batch UUID generation upfront to avoid repeated IO operations per span
- Make entity extraction pure computation by pre-generating all UUIDs
- Add OpenTelemetry tracing spans for pubsub and kafka batch processing with error tracking
- Add hashable dependency for HashMap-based deduplication
- Remove unnecessary threadDelay in test utilities
- Add strictness annotations to improve memory usage
These changes significantly reduce the algorithmic complexity and IO overhead of span processing while improving observability.
* use details and summary for presets dropdown.
* tiny query lib improvements
* improve the list better
* allow deletes remove and update list via htmx
* x
* reduce dom count of the facets elements.
* checkpoint
* fix not collapsing parent without children
* remove codemirrow
* enable source maps, update echarts, update tailwind, etc
* cleanup listeners and be more lazy in session reploay
* onboarding
* support for basic auth for open source users
* implement basic auth support
* introduce support for basic auth and support disabling the client side monitoring via monoscope-browser
* pass in the entire config to bodywrapper
* support leaving unused fields as optional
* make fmt
* add posthog
* configure external scripts via env variables
* make fmt
* fix compile errors
* replay events and publish/receive from/to pubsub
* pubsub for rrweb events works
* allowing CORS for all origins in OpenTelemetry
* fix: allow messages with project IDs
* session ui
* session replay modal
* towards fixing hls
* revert
* feat: save session files
* remove unused save file variants
* fetch and serve replay events in ui
* use background ctx
* remove amazonka dependency
* update yamr loc
* format
* fix completion issue on query editor
* swap agg and by
* checkpoint agg works
* checkpoint group by
* group by with functinoing binning ui
* checkpoint: figure out sort by input
* checkpoint
* checkpoint
* checkpoint
* checkpoint
* show and hide sort by limit and more correctly
* generating new sumamrize by syntax
* support sumamry by
* checkpoint
* checkpoint
* modify dash temps
* anomaly list
* page chart widgets
* stick to kql as much as possible
* sumarize
* sumarize
* sumarize
* sorta functional
* functioning aggs
* checpoint working editor. need to submit query and actually search
* working handleAddQuery impl
* ask ai btn
* llm search checkpoint before integrate langchain
* working ai search checkpoint
* chekcpoint not overlapping suggestions and can search
* checkpoint
* flatter schemas support
* prettier config
* use ? for ai queries
* checkpoint. TODO: include schema to ai search with a schema resolution endpoint
* move schema to backend
* widget fixes to explorer
* register popular queries