Hyperdrive, Workers - Hyperdrive exposes database connection pool size metrics

Summary

Cloudflare Hyperdrive now exposes detailed connection pool size metrics so engineers can self-diagnose connection contention. Metrics are available in the Cloudflare dashboard (Metrics tab → "Pool connections" chart) and via the GraphQL Analytics API dataset hyperdrivePoolSizesAdaptiveGroups.

Key Points

• Dashboard: new "Pool connections" chart on each Hyperdrive configuration shows Waiting clients, Open connections, and Pool size maximum. Use the location selector (airport codes) to drill down by location.
• GraphQL dataset: hyperdrivePoolSizesAdaptiveGroups exposes both averaged and peak values:
  • avg: currentPoolSize, availablePoolSlots, waitingClients
  • max: maxPoolSize, currentPoolSize, waitingClients
• Interpretation: spikes in waitingClients or sustained currentPoolSize near maxPoolSize indicate contention. If currentPoolSize regularly approaches maxPoolSize, contact Cloudflare to request a higher Hyperdrive connection limit.

Practical steps for engineers

• Add hyperdrivePoolSizesAdaptiveGroups queries to existing analytics workflows to alert on rising waitingClients or currentPoolSize thresholds.
• Use location drill-down to isolate affected POPs (airport codes) and correlate with app traffic patterns.
• If alerts trigger frequently, consider increasing connection limits with Cloudflare or optimizing application connection usage.

References

See Metrics and analytics and Connection pooling in Cloudflare docs for full schema and query examples.