Amazon Redshift Integration
Monitor WLM queue metrics, query execution plans, and cluster storage utilization across your Amazon Redshift warehouses. Get AI-powered query regression detection and predictive storage alerts.
How It Works
Connect via AWS CloudWatch & Redshift APIs
TigerOps bridges Amazon Redshift CloudWatch metrics and SVL/STL system table queries using an IAM role with read-only permissions. No agent installation inside your Redshift cluster is required.
Configure IAM and Cluster Access
Attach the TigerOps IAM policy to a cross-account role. TigerOps uses this role to pull CloudWatch metrics and connect to Redshift system tables via the Data API with a read-only database user.
Enable WLM Queue Monitoring
TigerOps queries stv_wlm_service_class_state and svl_query_summary to track WLM queue depths, concurrency slots used, and queue wait times per service class and user group.
Set Storage and Query Alerts
Define thresholds for cluster storage utilization, WLM queue depth, and query execution time SLOs. TigerOps correlates query plan regressions with statistics staleness and sort key effectiveness.
What You Get Out of the Box
WLM Queue Depth & Concurrency
Per-service-class queue depth, concurrency slot utilization, and queue wait time histograms. TigerOps alerts when queries are queuing in WLM and identifies which user groups are consuming the most slots.
Query Execution Plan Analysis
Top queries by execution time, disk spill events, nested loop operations, and broadcast operations from SVL_QUERY_SUMMARY and SVL_QUERY_REPORT. Surface inefficient plans before they degrade dashboard performance.
Cluster Storage Utilization
Per-node storage used and free, table skew ratios, and vacuum effectiveness. TigerOps predicts when a node will reach storage limits based on current growth rates and alerts before disk pressure causes query failures.
Node & Slice Health
Per-node CPU utilization, network throughput, and disk I/O from CloudWatch. Detect uneven workload distribution across nodes and identify slice-level skew in distributed query execution.
VACUUM & ANALYZE Tracking
Track table vacuum progress, unsorted row percentages, statistics staleness, and ANALYZE execution frequency. Alert when high unsorted ratios indicate that tables need vacuuming to maintain sort key efficiency.
AI Query Regression Detection
TigerOps AI compares query execution metrics before and after Redshift cluster resizes, maintenance windows, or statistics updates to automatically identify which queries regressed and why.
TigerOps Config for Amazon Redshift
Connect TigerOps to Redshift via AWS IAM cross-account role — no agent in your cluster required.
# TigerOps Amazon Redshift integration config
# Place at /etc/tigerops/conf.d/redshift.yaml
integrations:
- name: amazon-redshift
type: redshift
config:
aws:
region: us-east-1
# Cross-account IAM role assumed by TigerOps
role_arn: "arn:aws:iam::123456789012:role/TigerOpsRedshiftMonitor"
external_id: "${TIGEROPS_EXTERNAL_ID}"
# Cluster(s) to monitor
clusters:
- cluster_identifier: my-redshift-cluster
database: analytics
db_user: tigerops_monitor
# Use Redshift Data API (no persistent connection)
use_data_api: true
# Metrics to collect
metrics:
cloudwatch:
- CPUUtilization
- PercentageDiskSpaceUsed
- ReadIOPS
- WriteIOPS
- NetworkReceiveThroughput
- NetworkTransmitThroughput
system_tables:
- wlm_queue_stats # stv_wlm_service_class_state
- query_summary # svl_query_summary (top 50 by time)
- table_stats # svv_table_info (skew, unsorted)
- vacuum_progress # svv_vacuum_progress
# WLM service class mapping (optional)
wlm_service_classes:
"1": default
"5": etl
"6": reporting
"100": superuser
# Data API polling interval
system_table_interval: 60s
remote_write:
endpoint: https://ingest.atatus.net/api/v1/write
bearer_token: "${TIGEROPS_API_KEY}"Common Questions
Does TigerOps support Amazon Redshift Serverless?
Yes. TigerOps supports both Redshift provisioned clusters and Redshift Serverless. For Serverless, TigerOps collects RPU utilization, query duration, data scanned per query, and workgroup-level capacity metrics via CloudWatch.
What IAM permissions does TigerOps require?
TigerOps requires a cross-account IAM role with cloudwatch:GetMetricStatistics for Redshift namespaces, redshift-data:ExecuteStatement and redshift-data:GetStatementResult for system table queries, and redshift:DescribeClusters for cluster inventory. No superuser database access is needed.
How does TigerOps monitor WLM without a persistent Redshift connection?
TigerOps uses the Redshift Data API to execute system table queries on a schedule (every 60 seconds by default). This avoids holding open connections and works with both IAM authentication and database user credentials.
Can TigerOps detect table skew in Redshift?
Yes. TigerOps queries SVV_TABLE_INFO to collect per-table skew ratios (max slice rows / average slice rows) and flags tables where high skew indicates a poor distribution key choice. These are surfaced as optimization recommendations alongside query latency metrics.
How are Redshift maintenance windows handled in alerting?
TigerOps imports your Redshift cluster maintenance windows from the DescribeClusters API and automatically suppresses expected restart alerts during those windows. Post-maintenance, TigerOps checks for query plan regressions caused by statistics resets.
Stop Reacting to Redshift Slowdowns After the Fact
WLM queue monitoring, query plan analysis, and AI regression detection for Amazon Redshift. Connect in minutes via IAM.