Question 1

Which databases does ClusterEye support?

Accepted Answer

ClusterEye supports 14 database engines: PostgreSQL, Oracle Database, Microsoft SQL Server, MongoDB, MySQL, MariaDB, ClickHouse, Redis, Elasticsearch (and OpenSearch), Azure SQL Database/Managed Instance, Azure Database for PostgreSQL, AWS RDS for PostgreSQL and SQL Server, Huawei Cloud RDS for PostgreSQL. Each engine has its own engine-specific metrics, ready-made dashboards and query analysis.

Question 2

Do I have to install an agent on every database?

Accepted Answer

No. PostgreSQL, Oracle, SQL Server, MongoDB, MySQL/MariaDB and ClickHouse are monitored with an agent that runs on the database host, which is what makes host-level metrics, log analysis and configuration drift possible. Redis, Elasticsearch, Azure SQL, Azure PostgreSQL, AWS RDS (PostgreSQL and SQL Server) and Huawei Cloud RDS for PostgreSQL are monitored agentlessly: the connection is made directly from the ClusterEye API server and nothing is installed on the target.

Question 3

Does Oracle monitoring require a Diagnostics or Tuning Pack licence?

Accepted Answer

No. ClusterEye never queries the DBA_HIST_* (AWR) views or V$ACTIVE_SESSION_HISTORY on Oracle, so no Diagnostics Pack or Tuning Pack licence is required. Wait event analysis, top SQL, blocking trees and tablespace forecasting are all derived from V$ and DBA_ views that carry no additional licence.

Question 4

What privileges does the monitoring account need?

Accepted Answer

The baseline is read-only on every engine and is safe to grant in production: pg_monitor on PostgreSQL, SELECT_CATALOG_ROLE plus SELECT ANY DICTIONARY on Oracle, VIEW SERVER STATE and VIEW ANY DEFINITION on SQL Server, clusterMonitor and readAnyDatabase on MongoDB. Action features such as killing a session or forcing a plan are unlocked by separate, optional grants; without them the feature degrades to recommend-only.

Question 5

Which firewall ports do I need to open?

Accepted Answer

For agent-based engines a single rule is enough: outbound TCP from the database host to port 443/tcp on the ClusterEye server. The agent only dials out and never opens a listening port on the monitored host. For agentless engines the direction is reversed — the ClusterEye API server needs to reach the target port (Redis 6379, Elasticsearch 9200, Azure SQL ve AWS RDS SQL Server 1433, Azure/AWS/Huawei PostgreSQL 5432).

Question 6

Can ClusterEye run on-premises or in an air-gapped environment?

Accepted Answer

Yes. ClusterEye can run entirely inside your own infrastructure: a single-command standalone install on bare-metal Linux, or on Kubernetes. In air-gapped environments the agent binary can be copied in by hand, so no internet access is required to install or operate it. Your data never leaves your own servers.

14 engines, one pane.

Metrics tailored to each engine

Frequently asked questions

Don't see your engine in the list?