Database outages often stem from subtle configuration issues like problematic prefetch settings. Current monitoring tools don't catch these until it's too late.
Build a Postgres-specific monitor that tracks unusual query patterns and warns about risky configurations. Include automated remediation suggestions.
Cloud providers and enterprises would pay to prevent costly downtime. Offer both self-hosted and SaaS versions.
MVP could be a set of curated Grafana dashboards with specific alerts.
The risk is convincing ops teams to adopt yet another monitoring tool.