Changelog
Every update, documented
New features, improvements, and bug fixes — shipped continuously.
NewImprovedFixedBreaking
v0.9.3Latest
AI anomaly detection and dashboard builder improvements
- NewAI-powered anomaly detection with automatic root-cause suggestions
- NewDashboard builder: drag-and-drop panel layout with 6 new widget types
- NewOpenTelemetry Collector 0.98 support with new receiver types
- ImprovedPromQL editor now has autocomplete, syntax highlighting, and error hints
- ImprovedTrace waterfall view performance — up to 4× faster rendering for deep traces
- FixedDashboard snapshot links no longer expire after 24 hours
- FixedAlert rule editor now correctly validates PromQL expressions before saving
v0.9.2
Log pattern clustering and improved alert routing
- NewLog pattern clustering automatically groups similar log lines to surface anomalies
- NewAlert routing rules: send different severity alerts to different channels
- NewService dependency map in the Traces overview
- ImprovedLog search is now up to 10× faster with new inverted index
- ImprovedAPI keys now show last-used timestamp and request count
- FixedFixed a race condition where live tail could drop log lines under high throughput
v0.9.1
SLO tracking and multi-window burn rate alerts
- NewSLO management: define error budgets and track burn rate in real time
- NewMulti-window burn rate alerts (1h + 6h Alertmanager-compatible rules)
- NewExemplars: click any metric data point to jump to a representative trace
- ImprovedMetrics API now returns consistent timestamps across multi-series responses
- FixedFixed NaN values in histogram quantile calculations for sparse buckets
- FixedTeam invitations now correctly honour the invited user's timezone preference
v0.9.0
Major release: Logs GA and new data model
- NewLogs is now generally available on all plans, including the free tier
- NewUnified data model: logs, metrics, and traces share the same resource attributes
- NewKafka adapter for ingesting logs from existing Kafka pipelines
- BreakingThe v1 Metrics API has been removed. Migrate to v2 before upgrading.
- ImprovedIngestion pipeline latency reduced from ~800ms to ~120ms p50
- FixedOpenSearch storage backend now correctly handles index rollover at midnight UTC
v0.8.5
Performance and reliability improvements
- ImprovedReduced cold-start latency for serverless receiver by 60%
- ImprovedClickHouse schema migration to ReplacingMergeTree for deduplication
- FixedFixed memory leak in the metric aggregation pipeline under high cardinality
- FixedOTLP receiver now correctly handles batch requests with mixed compression
Looking for older releases?
View full history on GitHub