ShellYard

Health dashboards

Live engine health on a user-set cadence with pause / resume. Per-engine breakouts and Magellan presets for 'Spot anomalies' and 'Capacity check'.

Every DB workspace has a Health tab next to Query. It polls the engine on a user-set cadence (default 5s) with Pause / Resume / Refresh-now controls and renders results as tone-graded metric cards (ok / warn / error) plus tabular detail.

Per-engine reads

Postgres

  • Version, uptime, role (primary / standby)
  • Connection pool vs max_connections (color-graded)
  • Buffer cache hit %, commit / rollback ratio
  • Top databases by size
  • Active queries from pg_stat_activity
  • pg_stat_replication lag

MySQL

  • SHOW GLOBAL STATUS — threads vs max_connections, traffic counters
  • InnoDB buffer pool hit ratio
  • Top schemas by size
  • SHOW PROCESSLIST (excluding our own connection)
  • SHOW SLAVE STATUS for replicas

Microsoft SQL Server

  • sys.dm_exec_requests — active queries with wait type and blocking session id
  • sys.dm_os_wait_stats — cumulative wait totals; top wait types surfaced
  • sys.dm_db_index_usage_stats — index seeks vs scans, useful for spotting missing indexes
  • sys.master_files — database size and growth per file
  • Replica state via sys.dm_hadr_database_replica_states when AGs are configured

SQLite

  • File size, page geometry
  • Free pages / fragmentation %
  • Journal mode (WAL / DELETE / …)
  • Per-table row counts
  • PRAGMA quick_check integrity probe

Redis

  • INFO sections — server, clients, memory vs maxmemory, stats with hit ratio, persistence, replication
  • DB keyspace per logical DB
  • SLOWLOG GET 10

MongoDB

  • serverStatus — connections, opcounters, memory, WiredTiger cache hit ratio
  • Replica-set state
  • dbStats per database
  • currentOp active operations

Magellan attach

The Magellan button on the Health header ships the flattened snapshot — sections, metrics, tone hints, slow-query tables — as context. Two presets on the button:

  • Spot anomalies — flags any metric in a warn or error band and explains likely causes
  • Capacity check — projects when current trends hit the engine’s hard limits (max_connections, maxmemory, replication lag thresholds, etc.)

See DB Health context for the context payload details.