Engine Management
LakeOps provides a unified view of all query engines connected to your Iceberg lake. Monitor health, compare performance, and add new engines — without engine-specific scripts or duplicate tooling.
Why unified engine management?
Modern data lakes connect multiple query engines for different workloads: Trino for interactive analytics, AWS Athena for serverless queries, Snowflake for BI dashboards, DuckDB for local development. Without a unified view, teams end up with fragmented monitoring, inconsistent configurations, and no way to compare engines objectively.
LakeOps solves this by providing a single control plane where you can see every engine's status, usage, cost, and health — and take action from one place.
Supported engines
LakeOps supports the following query engines out of the box:
Additional engines can be connected via the Add Engine wizard with custom connection parameters.
Query engines screen
Navigate to Engines > Overview in the sidebar. The screen provides a complete view of your engine fleet.
Summary cards
Four stat cards at the top show fleet-level metrics:
Quick links
Below the stat cards, three action tiles provide shortcuts:
- •Compare engines — side-by-side cost, latency, and volume comparison
- •Engine health — uptime, incidents, and resource signals per engine
- •Add engine — guided wizard for credentials and connection tests
Engine directory
The main area shows engine cards in a grid layout. Each card displays:
| Field | Description |
|---|---|
| Engine name | Name and type with brand icon |
| Status | Active (green), Inactive (gray), or Maintenance (amber) |
| Queries | Total queries routed to this engine |
| Avg runtime | Average query execution time |
| Cost / query | Average cost per query execution |
| Last used | How recently the engine processed a query |
Use the search bar and status filter to find specific engines quickly.
Engines Compare
Navigate to Engines > Compare to benchmark engines side by side. The compare view shows:
- •Query success rate — percentage of queries completed without errors
- •Average runtime — mean query execution time per engine
- •Cost per query — average execution cost for direct comparison
- •Total queries — query volume processed per engine
- •Data scanned — total data read during query execution
Toggle engines from the engine tiles to populate the comparison table. A cost vs. performance visualization helps identify the most efficient engine for each workload.
Engine Health
Navigate to Engines > Health for real-time health monitoring. The health view shows each engine with:
- •Status indicator — Healthy, Inactive, or Maintenance
- •Uptime percentage — availability over the monitoring period
- •Response time — current average latency per engine
- •Resource utilization — CPU, Memory, Disk I/O, and Network usage bars per engine
- •CHECK NOW button — trigger an immediate health check
Adding a new engine
Navigate to Engines > Add Engine in the sidebar. The guided wizard walks you through four steps:
Engine statuses
| Status | Meaning | Routing behavior |
|---|---|---|
| Active | Engine is connected, healthy, and accepting queries | Eligible for query routing |
| Inactive | Engine is registered but not currently connected | Excluded from routing until reactivated |
| Maintenance | Engine is temporarily unavailable (planned downtime) | Excluded from routing; failover triggers for in-flight queries |
Engines & routing
Engines are the foundation of query routing. Each routing group references one or more engines. When you add or remove an engine, routing groups that reference it are updated automatically. If an engine enters Maintenance or Inactive status, the routing layer excludes it and fails over to remaining engines in the group.