Agents and modules generated by the plugin

The plugin will create an agent for each target database. That agent will contain the following modules

If engine_uptime is activated:

restart_detectionrestart\_detection

It will be 0 if an unexpected restart has been detected, and 1 if this is not the case. When a server restarts unexpectedly, there may be an interruption in access to the database and potentially lost transactions or data not saved correctly.

Si esta activado query_stats:

queries: select	**Number of SELECT queriesqueries**. SELECT queries are used to retrieve data from the database. Monitoring SELECT queries allows you to assess query efficiency and index optimization. By identifying slow or inefficient SELECT queries, you can take steps to improve their performance, such as adding appropriate indexes, optimizing the queries, or adjusting server configuration.
queriers: update	**Number of UPDATE queriesqueries**. UPDATE queries are used to modify existing data in the database. Monitoring UPDATE queries is important to assess the frequency and efficiency of data updates. You can identify UPDATE queries that affect a large number of rows or have a significant impact on server performance. This allows you to optimize queries, review table structure, or take steps to reduce the load generated by updates.
queries: delete	**Number of DELETE queriesqueries**. DELETE queries are used to delete data from the database. Monitoring DELETE queries is useful for evaluating the frequency and efficiency of data deletions. You can identify DELETE queries that affect a large number of rows or that have a significant impact on server performance. This allows you to optimize queries, review table structure, or take steps to reduce the load generated by deletes.
queries: insert	**Number of INSERT queriesqueries**. INSERT queries are used to insert new data into the database. Monitoring INSERT queries allows you to assess the frequency and efficiency of data inserts. You can identify INSERT queries that are placing a high load on the server or could be causing performance issues. This allows you to optimize queries, review table structure, or consider lazy insert strategies to improve performance in high-concurrency environments.

If analyze_connections is activated:

session usage

**Number of current connectionsconnections** with respect to the total of maximum connections. Monitoring session usage in SQL Server is important to optimize performance, identify blocking issues, improve security and auditing, and efficiently plan server resources.

If cache_stats is enabled:

cache hit ratio (dictionary)	**The cache ratio of the dictionary.** "Cache Hit Ratio (Dictionary)" refers to the ratio of Oracle dictionary data access requests that are resolved using cached data compared to requests that require disk access. The Oracle dictionary contains information about the database structure, objects, metadata, and other important details. A high hit ratio indicates that most dictionary access requests are resolved using data in the cache, which improves system performance by avoiding disk access.
cache hit ratio (library)	**The cache ratio of the library.** "Cache Hit Ratio (Library)" refers to the ratio of access requests to Oracle procedures, functions, and packages stored in the shared library cache. The shared library cache stores the compiled code of frequently executed programs and SQL queries. A high hit ratio indicates that most requests are resolved using cached code, which avoids the need for recompilation and improves performance by reducing execution time.
cache hit ratio (buffer)	**The buffer cache ratio.** "Cache Hit Ratio (Buffer)" refers to the proportion of requests to access blocks of data stored in Oracle's buffer cache that are resolved using data in the cache compared to requests that require access to disk. The buffer cache stores blocks of data that are frequently read or modified. A high hit ratio indicates that most data access requests are resolved using blocks stored in the buffer cache, which reduces the need for disk access and improves overall system performance.

If check_tablespaces is enabled:

tablespace <table name> free

**Percentage of usage in GBGB**. It is important to monitor "tablespace free" in Oracle to keep track of the amount of space available in the tablespaces. This allows you to proactively manage storage space growth and avoid insufficient space issues that could impact database functionality. By monitoring tablespace free, you can take actions such as adding more storage space or making adjustments to configurations to ensure there is enough space available.

tablespace <table name> status

**Status of the table, 1 if it is online and 0 if notnot**. Monitoring tablespace status in Oracle is essential to evaluate the health and status of tablespaces. Provides information about the structural integrity of tablespaces, including the existence of errors or corruption issues. By monitoring tablespace status, you can detect problems early and take action to resolve them, such as restoring from backups or performing recovery tasks. This ensures database continuity and stability, preventing data loss or performance degradation due to tablespace issues.

If fragmentation_ratio is activated:

fragmentation ratio

**The fragmentation ratioratio**. The Fragmentation Ratio in Oracle Database is a value that indicates the amount of space wasted due to fragmentation in a data structure. A high Fragmentation Ratio indicates greater fragmentation and wasted space, which can impact system performance.

The plugin will also create a module for each custom query defined in the configuration file.

---

Multitenant Monitoring (CDB/PDB)

If multitenant is enabled (1 — enabled, 0 — disabled; default: 0), the plugin discovers Pluggable Databases (PDBs) within the Container Database (CDB) and collects the same performance metrics for each PDB (uptime, query stats, cache hit ratio, tablespaces, connections, fragmentation). Only applicable to Oracle multitenant (CDB/PDB) architectures; in non-CDB environments it will simply emit a harmless warning.

Behavior depends on the agent_per_pdb token:

agent_per_pdb disabled (default)

All PDB modules are added to the main CDB instance agent. Module names are prefixed with PDB <name>, for example:

PDB APP1 restart_detection	Restart detection for PDB APP1.
PDB APP1 queries: select	SELECT queries on PDB APP1.
PDB APP1 cache hit ratio (buffer)	Buffer cache hit ratio for PDB APP1.
PDB APP1 tablespace USERS free	Free space of tablespace USERS in PDB APP1.

agent_per_pdb enabled

The plugin creates an independent agent for each discovered PDB, named <main_agent> - PDB <name> with address host:port/pdb_name. Each PDB agent contains:

• A connection module (Oracle PDB <name> connection)
• All applicable statistics modules (uptime, queries, cache, tablespaces, connections, fragmentation), same as the main CDB agent
• Custom thresholds are also applied to each PDB agent's modules

PDBs not in READ WRITE mode are skipped with a warning.

Note: agent_per_pdb only takes effect when multitenant enabled.