Live Observation
The Live Observation panel provides a real-time view of cluster activity and health. Unlike the other Redshift Health views which analyze historical data, this panel queries Redshift system views for current state and auto-refreshes to keep the display current.
Live Queries
The Live Queries panel shows all queries currently executing on the cluster. Data source: stv_wlm_query_state, which reflects the current WLM state without any caching.
| Column | Description |
|---|---|
| Query ID | Redshift internal query identifier |
| Exec Time (s) | Seconds the query has been executing. Highlighted red if > 300s. |
| WLM Queue | The WLM queue slot the query is occupying |
| User | Database user running the query |
| Database | Target database for this query |
| Query Text | First 120 characters of the SQL text. Click to see the full query. |
The panel auto-refreshes every 30 seconds. A countdown timer shows the next refresh. Click "Refresh Now" to force an immediate update.
Node Storage
Per-node storage utilization bars show used vs. available disk capacity for each cluster node. Data source: stv_slices joined with stv_node_storage_capacity.
Node type is detected by matching the node's reported capacity against known node type capacities:
| Node Type | Capacity (MB) |
|---|---|
| dc1.large | 190,633 MB |
| dc2.large (NVMe) | 190,633 MB |
| dc2.8xlarge | 760,956 MB |
| ra3.xlplus | 952,455 MB |
| ra3.4xlarge | 3,339,176 MB (1 partition) |
| ra3.16xlarge | 3,339,176 MB (4 partitions) |
Application Performance
A 7-day rolling line chart showing the daily average query execution time per application. Data source: sys_query_history joined with sys_connection_log to map queries to their application names.
Each line represents one application. Use the legend to toggle individual applications on or off. This view helps identify which application's queries are trending slower over time.
Cluster Health Score
A composite score (0–100) computed from three sub-scores, each weighted equally:
| Component | Metric Used | Healthy Threshold |
|---|---|---|
| Vacuum Health | Average unsorted % across all user tables | < 10% |
| Query Health | Percentage of queries in last 24h that had disk spill | < 5% |
| WLM Health | Ratio of queue wait time to total query time in last 24h | < 15% |
Auto-Refresh