Monitor databases
Add a PostgreSQL, MySQL/MariaDB, SQL Server, or Oracle database with a read-only user and review its login, space, sessions, locks, replication, and backups.
On this page
ARGUNIS checks your databases' connection and login, space, sessions, locks, long transactions, replication, backups, and locked or expiring users, always with a read-only user.
Before you begin
You need permission to operate in Infrastructure.
Create a read-only user in the database. In Infrastructure → Databases, the Read-only user script button provides the script for each engine so you can copy or download it.
Save that user as a database credential in Settings → Monitoring → Credentials.
If the database is on a private network, you need a monitoring probe, an agent, or a link on that network.
Steps
Go to Infrastructure → Databases.
Click Add database.
Choose the Engine: PostgreSQL, MySQL / MariaDB, SQL Server, Oracle, ClickHouse, Vertica, Db2, Informix, MongoDB, Redis / Valkey, Cassandra / ScyllaDB, or Elasticsearch / OpenSearch.
Engine
What is checked besides "responds" and connection time
ClickHouse (HTTP 8123/8443)
Disk per volume, long queries, lagging or read-only replicas, last backup
Vertica
Space per node, sessions, long queries, down nodes, last backup
Db2 (LUW)
Tablespaces, sessions, locks, long transactions, HADR, last backup
Informix (via DRDA)
Dbspaces, sessions, locks, last level-0 archive
MongoDB
Connections against the maximum, space per database, long operations, replica set and its lag
Redis / Valkey
Memory against
maxmemory, clients, replicas and their lag, last RDB saveCassandra / ScyllaDB
Nodes up or down, schema agreement, connected clients
Elasticsearch / OpenSearch
Cluster health (red = alarm), unassigned shards, disk per node, last snapshot
Db2 and Informix use the official IBM client, only on 64-bit x86 hosts. On an agent, the administrator installs it once with argunis-agent install-ibm-db (it requires npm, Python 3, a C++ compiler, and libxml2); in the cloud and on the monitoring probe it is already included. Informix connects via DRDA: the server needs a drsoctcp listener (for example, port 9089).
Enter the Name you'll see it under.
In Connection:
Address (host or VIP).
Port (optional): by default 5432 (PostgreSQL), 3306 (MySQL), 1433 (SQL Server), and 1521 (Oracle).
The Database (PostgreSQL, MySQL, SQL Server) or, in Oracle, the Service name or the SID.
MongoDB Atlas (or another cluster with an SRV record): select SRV address (mongodb+srv) and enter the cluster name (for example
cluster0.abcd.mongodb.net), or paste themongodb+srv://…string that Atlas provides (ARGUNIS takes only the name; the user and password go in the credential). The nodes come from DNS on every check, and the connection always uses TLS. In Atlas, add the outbound IP of wherever the check runs from (the monitoring probe on your network or ARGUNIS in the cloud) to its access list (Network Access). The free Atlas M0 plan doesn't provide all metrics (for example, long operations).Nodes behind the load balancer (optional): one per line,
hostorhost:port. The VIP and each node are checked separately (up to 8).
In Access and location:
Vault credential: the read-only user. If it's missing, use + Create credential.
Run from: the ARGUNIS cloud, a link, an agent, or an active monitoring probe.
Interval (seconds): from 60 to 3600 (300 by default).
Inventory group and Server that hosts it (optional).
Click Add and check.
What you see in the list
At the top, a summary per engine with how many databases are up, down, degraded, or without data. In the table, each database with its status (Up, Degraded, Down, or No data), engine and version, address, login time, fullest space, sessions, blocked sessions, and alarms.
On each row you can Check now, Edit, or Stop monitoring.
The database page
Click the database name to open its page. It has Check now, Configure, and these tabs:
Tab | What it shows |
|---|---|
Summary | The connection points (the VIP or main address and each node) and the database details. |
Monitors | What is checked on the database, with its status and Run. |
Charts | Login, sessions (total and active), worst space, cache hits, and replication lag over the period. |
Queries and sessions | Sessions and locks (who is blocking whom) and Long transactions. |
Space and backups | Tablespaces, databases, and files with their growth over 24 hours and 7 days, replication, backups, and locked or expiring users. |
Events, Incidents, Relationships, and Comments | Same as on a host page. |
If the read-only user couldn't read a section, the page says so in Notes.
Configure the database
On the page, Configure has these tabs:
General: environment, site, and what the database does when it fails (Creates incident and resolver group).
Connection and credentials: engine, address, nodes, database, credential, and where it's checked from. Change them with Edit.
Thresholds: the current thresholds. See Adjust the alert thresholds.
Monitors and Tests and macros: same as on a host. See Associate monitoring tests and use tags.
Managed databases from your cloud services appear at the end of the list, in Cloud services.