Operations and observability
Once your game is running on Heroic Cloud, you need visibility into how it’s performing and tools to act on what you see. This section covers the observability and operational tooling available to you.
Overview #
Heroic Cloud provides built-in monitoring, logging, and data export for every Nakama deployment. For studios that need deeper integration with their own observability stack, paid add-ons extend these capabilities with continuous log shipping, Prometheus-compatible metric endpoints, read-only database replicas, and application profiling.
What’s included with every deployment #
Every Nakama deployment comes with the following at no additional cost:
- Graphs showing time-series performance data: load balancer request counts, CPU and memory utilization per node, database CPU, and database query load. See Dashboard metrics and logs.
- Logs with full-text search, severity filtering, date range selection, and one-time export. All logs are in UTC. See Dashboard metrics and logs.
- Top database queries surfacing the most expensive SQL queries with an impact indicator to identify optimization targets.
- Audit trail tracking every user action on the deployment: who deployed, who scaled, who changed configuration. See Audit log.
- Data export for downloading a complete PostgreSQL snapshot of your deployment’s database at any time. See Data exporting.
Additional add-ons #
For studios that run their own observability platforms or need capabilities beyond the built-in tools:
- Log exporting ships logs to an S3 bucket in near real-time (20-30 second delay) as JSON. Ingest into Elasticsearch, Splunk, Datadog, Grafana Loki, or any platform that reads from S3. See Log exporting.
- Metric exporting provides a Prometheus-conforming endpoint you scrape every minute. Includes Nakama application metrics and any custom metrics you emit from your game module. Plot in Grafana or any compatible tool. See Metric exporting.
- Database replica gives you a read-only database accessible over the internet for querying production data without impacting performance. See Additional add-ons.
- Profiling provides application-level profiling for your Nakama instances. See Additional add-ons.
To enable any add-on, contact Heroic Labs (support@heroiclabs.com). See Billing for how add-on charges work.
When to use what #
| Need | Tool | Cost |
|---|---|---|
| Quick health check or debugging | Built-in graphs and logs | Included |
| Identify slow queries | Top database queries | Included |
| Track who changed what | Audit log | Included |
| Export your data as a one-off in SQL format. This is not for backing up or migration. | Data export | Included |
| Export Nakama logs in near real-time to an S3 bucket, per deployment | Log exporting add-on | Optionally enabled |
| Scrape Heroic Cloud metrics in Prometheus format | Metric exporting add-on | Optionally enabled |
| Read-only database access on a real-time replicated database | Database replica add-on | Optionally enabled |
| CPU, Memory, and Mutex Contention continuous profiling | Profiling add-on | Optionally enabled |
Disaster recovery #
Heroic Cloud automatically backs up your database daily and replicates data across availability zones for high availability deployments. See Disaster recovery and backups.
Load testing #
Before going live, validate your deployment’s performance under realistic load. Development instances can’t be used for load testing. See Load testing guide.
See also #
- Scaling for resource tiers and when to scale.
- Nakama deployments for the Graphs, Logs, and Data Export tabs.
