Operations

This section covers GQLDB administration and operations, including the tasks an operator does to keep a running database healthy, observable, and recoverable. The focus is how to operate the database, distinct from how to query it.

What's in Scope

• Install & deploy: installing GQLDB Community or Commercial Edition, sizing a host, picking a deployment topology (single-node, 3-node HA, 2-data + witness), and bringing the database up for the first time. See Database Installation and Deployment Topologies.
• Operate a cluster: observing leader / follower state, replication lag, planned failover, adding and removing voters, and recovering from quorum loss. See Clustering.
• Cloud & marketplace: running GQLDB as a managed service on Ultipa Cloud (DBaaS) or as a self-managed AMI from AWS Marketplace — what each path handles for you, what you still run, and when to pick which. See Cloud Deployments.
• Inspect a running database: version, license, loaded plugins, per-graph statistics, schema, and current-graph context. See Database Info.
• Backup & restore: taking hot full and incremental backups at graph or database scope, listing and verifying archives, dropping catalog entries, restoring with or without overwrite. See Backup & Restore.
• Monitoring: logs, Prometheus-style metrics, gRPC health checks, HA telemetry, query monitoring, memory-pressure watermark, alert-worthy signals. See Monitoring.
• Performance: indexes, compute engine, edge _id, memory and WAL flags, statistics freshness, storage hardware, HA read routing, bulk import. A catalog of levers with links to canonical references. See Performance.

What's NOT in This Section

Pages here are operator-focused: when to run a command, what to watch for, what to do if it fails, not the full syntax reference. Each page cross-links to the underlying GQL statement page for the canonical grammar.