Docs

Platform Guide

This guide covers everything you need to know about using the DBA SaaS platform effectively -- from navigating the dashboard to configuring advanced alert rules.

Dashboard Overview

The main dashboard is your central hub for monitoring all connected PostgreSQL databases. It provides an at-a-glance view of your entire infrastructure.

Database List

The left panel displays all connected databases grouped by organization. Each entry shows:

  • Health indicator — a color-coded dot (green, yellow, red) reflecting the current state
  • Database name and host address
  • PostgreSQL version
  • Active connections count
  • Last metric received timestamp

Quick Stats

The top bar shows aggregated statistics across all your databases:

  • Total databases monitored
  • Open issues count by severity
  • Average query response time
  • Overall cache hit ratio

Monitoring & Metrics

DBA SaaS collects a comprehensive set of metrics from your PostgreSQL instances. All metrics are gathered by the lightweight agent with minimal overhead.

Collected Metrics

  • Connections — active, idle, idle in transaction, waiting connections; connection pool utilization
  • Query Performance — slow queries, query execution times, queries per second, top queries by total time
  • Replication — replication lag, WAL generation rate, replica status, slot activity
  • Disk & I/O — table and index sizes, disk usage, read/write IOPS, temporary file usage
  • CPU & Memory — server CPU utilization, memory usage, swap activity
  • Cache Hit Ratio — shared buffer hit rate, index hit rate, toast hit rate
  • Vacuum & Autovacuum — last vacuum timestamps, dead tuple counts, autovacuum activity
  • Locks — lock waits, deadlocks, long-running lock holders

Data Retention

Metric retention varies by plan:

  • Free — 7 days of full-resolution data
  • Pro — 30 days full-resolution, 12 months aggregated
  • Enterprise — custom retention policies, up to unlimited

Issue Management

DBA SaaS continuously analyzes your metrics and automatically detects issues that need attention. Issues are surfaced in the dashboard and via configured notification channels.

How Issues Are Detected

The AI engine evaluates incoming metrics against baseline patterns, known anti-patterns, and configurable thresholds. When an anomaly is detected, an issue is created with:

  • A clear description of the problem
  • The affected database and relevant metrics
  • A recommended resolution action
  • Historical context showing when the problem started

Severity Levels

Each issue is assigned one of three severity levels:

  • Critical — immediate action required; the database is experiencing or will soon experience downtime, data loss, or severe performance degradation (e.g. replication stopped, disk full, connection exhaustion)
  • Warning — attention needed soon; a metric is trending in a concerning direction or has crossed a warning threshold (e.g. rising replication lag, low cache hit ratio, growing dead tuples)
  • Info — informational insight; an optimization opportunity or minor observation that does not require urgent action (e.g. unused index detected, configuration recommendation)

Resolving Issues

Issues can be resolved in two ways:

  • Auto-resolution — if the underlying metric returns to a healthy state, the issue is automatically closed with a note indicating when recovery occurred
  • Manual acknowledgment — team members can acknowledge an issue to indicate it is being worked on, or dismiss it if no action is needed

Alerts & Notifications

Configure alert rules to get notified about issues through the channels your team already uses.

Alert Rules

Create custom alert rules based on any collected metric. Each rule defines:

  • Metric — which metric to watch (e.g. connections.active, replication.lag_seconds)
  • Condition — threshold type (greater than, less than, equals) and value
  • Duration — how long the condition must persist before triggering (avoids false alarms from brief spikes)
  • Severity — the severity level to assign when triggered

Notification Channels

DBA SaaS supports the following notification channels:

  • Email — send alerts to individual addresses or distribution lists
  • Slack — post alerts to a Slack channel via incoming webhook
  • Webhook — send a JSON payload to any HTTP endpoint for custom integrations (PagerDuty, Opsgenie, custom scripts)

Alert Thresholds

DBA SaaS provides sensible default thresholds out of the box. You can customize these per database or globally:

  • Connection utilization > 80% triggers a Warning
  • Replication lag > 30 seconds triggers a Critical
  • Cache hit ratio < 95% triggers a Warning
  • Disk usage > 90% triggers a Critical

CLI Tools

For automation and scripting, DBA SaaS provides a command-line interface. The dbasaas CLI lets you manage databases, view metrics, and configure alerts directly from the terminal.

Common commands:

  • dbasaas databases list — list all monitored databases
  • dbasaas issues list --severity critical — view open critical issues
  • dbasaas alerts create — create a new alert rule
  • dbasaas metrics query --db mydb --metric connections.active — query recent metric values

The CLI is available on Linux and macOS and can be installed alongside the monitoring agent.