v0.2.0

[0.2.0] - 2026-03-26

Added

• Power commands: top (live ops), slow (query digest), explain (collection intelligence), ls (databases/collections), who (client dependency lookup)
• watch command for continuous anomaly detection with Slack notifications and delta reporting
• mcp command — MCP server mode with 7 agent-native tools (status, doctor, ls, top, slow, explain, who)
• users command for MongoDB user/role audit with unused account detection
• grep command to search collections and indexes by regex pattern
• diff command to compare two status snapshots with structured deltas
• snapshot command to save cluster state for offline analysis
• ci-init command to generate GitHub Actions and GitLab CI pipeline snippets
• --policy flag on status for policy-as-code enforcement (YAML rules)
• --baseline flag on status to suppress known conditions with auto-expiry
• --format sarif output on status and doctor for GitHub Code Scanning integration
• --quiet / -q global flag for machine consumption
• --verbose / -v global flag for debug logging
• Executive summary Grafana dashboard (4 key signals)
• systemd unit file and environment template in deploy/

Changed

• All logging migrated from log.Printf to structured slog (JSON in serve mode, text in CLI)
• DSN credentials masked in all log output via MaskDSN helper
• init command supports --with-policy to generate sample policy YAML

v0.1.0

[0.1.0] - 2026-03-26

Added

• Core CLI with five commands: serve, status, doctor, init, version
• 15 collectors producing 65 Prometheus metrics
• Replication monitoring: member state, lag per secondary, oplog window, election events
• Connection monitoring with exhaustion prediction (stateful forecasting)
• WiredTiger cache monitoring: bytes, dirty, eviction, read/write pages
• Operations: opcounters, currentOp (active/slow/longest), cursors, locks
• Data collectors: per-collection stats, per-database stats, network I/O
• Query regression detection via system.profile fingerprinting (unique)
• Unused index detection via $indexStats (unique)
• Election storm detection with stateful pattern analysis (unique)
• Sharding collectors: chunks per shard, balancer status, migration activity, shard key skew
• Built-in alerting: Telegram + webhook, 10 typed alerts with per-type cooldown
• Grafana anomaly annotations on spikes
• Multi-target support via comma-separated MONGO_DSN
• status --unhealthy filter and --format json on all commands
• ANCC-compliant doctor with 3-level exit codes (0=pass, 1=warn, 2=fail)
• Retry logic with exponential backoff for connectivity
• Helm chart with Deployment, Service, ServiceMonitor, PrometheusRule (7 alert rules)
• Grafana dashboard with 20 panels
• Dockerfile (distroless), CI workflow, release workflow with Homebrew tap automation
• ANCC-compliant SKILL.md with full interface declaration