Per-instance monitoring
Connections, storage, latency and replication lag for every instance, with metric history kept beyond what provider consoles retain at useful granularity.
Database Health
Database monitoring that finds the cost and the risk
Varcio Database Health monitors managed databases, principally Amazon RDS and Aurora with Azure and Google Cloud Cloud SQL paths, covering connections, storage, latency, replication lag and query performance, with forecasts, alerts and findings you can act on.
- orders-auroraConnections · latency · replicationHealthy
- billing-postgresCapacity forecastStorage pressure
- search-mysqlReplication lagLagging
- reports-sqlLow utilisationRight-size
14AWS database health detectors, from slow queries to table bloat
3 cloudsAWS RDS and Aurora, Google Cloud SQL and Azure pull paths
LongerMetric history retained beyond provider console retention
DemoA demo data mode to evaluate without a connected database
What it measures
Operational health and cost in one view
Databases are often the largest and least examined line on the bill, and the most sensitive to touch. Measured history makes optimising them a defensible decision.
Connections184Normal
Storage81%Forecast
Replica lag2 sNormal
Query-level evidence
Where you configure an optional direct connection, with credentials encrypted at rest, Varcio adds query sampling, a top-queries view and schema and configuration introspection.
Capacity forecasting
Storage and capacity trajectories are projected so exhaustion is anticipated instead of discovered.
Log diagnostics
Database logs are read for problems that metrics do not expose.
Alert rules
Configurable thresholds notify through your alert channel when a metric breaches.
Rule-driven recommendations
A rules engine evaluates the metric history and produces per-instance findings and recommendations.
Detection
Fourteen AWS database health detectors
Reliability, performance and cost findings from the same metric history.
14detectors
- Performance5
- Capacity4
- Reliability3
- Hygiene2
Performance
- Slow queries
- Expensive query plans
- Missing indexes
- Unused indexes
- Table bloat
Capacity
- Capacity forecast
- Storage pressure
- Connection saturation
- Low utilisation
Reliability
- Replication lag
- Deadlocks
- Anomalies
Hygiene
- Diagnostics logging disabled
- Security posture
Findings from Database Health flow into the same opportunity queue, approvals and Autopilot machinery as every other finding, and low-risk classes can be configured as auto-actions.
Getting value
From connection to savings
The highest-value findings come from query-level evidence, so the order matters.
- 01
Connect the cloud account
Run a sync to populate the instance inventory and confirm every production database appears.
- 02
Add direct connections for critical instances
This unlocks query-level sampling and introspection, where the best findings come from.
- 03
Work findings per instance
Over-provisioned instances and unused replicas are typically the largest cost items. Replication lag and connection saturation are the largest reliability items.
- 04
Act before exhaustion
Review storage forecasts and act before projected exhaustion rather than after.
- 05
Set alerts and auto-actions
Configure alert rules on the metrics that matter and automate only the finding classes that are safe to automate.
Plans
Part of the Workloads section
Database Health is available on Pro and above, alongside AI Infrastructure, API Monitoring and Kubernetes.
- Database Health
- AI Infrastructure
- API Monitoring
- Kubernetes
- Service Monitoring
- Real Users
AI Infrastructure
SageMaker, Bedrock, Vertex AI and Azure OpenAI findings and actions.
API Monitoring
The invoice’s view of your API layer, including the cost of failed requests.
Kubernetes
Cost by namespace, deployment and team, with rightsizing and idle-pod findings.
FAQ
Questions, answered
Which databases does Database Health support?
Principally Amazon RDS and Aurora, with pull paths for Google Cloud SQL and Azure. It reads metrics from the provider’s monitoring service and can optionally connect directly for query-level evidence.
Do I need a direct database connection?
No. Metric-based monitoring works from your connected cloud account. A direct connection, with credentials encrypted at rest, adds query sampling, a top-queries view and schema introspection.
Can it help reduce database cost?
Yes. Over-provisioned instances and unused replicas are typically the largest cost findings, and measured metric history, query evidence and capacity forecasts make rightsizing a defensible decision rather than a guess.
Can I try it without connecting a database?
Yes. A demo data mode lets you evaluate the module before connecting anything.
Which plans include it?
Pro, Business and Enterprise.
Keep reading
Related reading
AWS Cost Optimization Checklist: 30 Checks Across Compute, Storage, Network and Data
A working AWS cost optimization checklist: 30 checks grouped by compute, storage, network, databases and governance, with the pricing facts that tell you which ones matter most.
Read the articleHidden AWS Savings Opportunities: 12 Places Your Bill Is Leaking in 2026
NAT gateways, idle public IPs, gp2 volumes, forgotten snapshots, log retention and more. Twelve places AWS bills quietly leak money, with the pricing facts to size each one.
Read the articleFind out what your databases are really costing
Connect a cloud account and see database findings from your first sync. Free trial, no credit card required.