Oracle

Agent Check

Overview

Get metrics from Oracle Database servers in real time to visualize and monitor availability and performance.

Setup

Installation

Prerequisite

To use the Oracle integration, either install the Oracle Instant Client libraries, or download the Oracle JDBC Driver. Due to licensing restrictions, these libraries are not included in the Datadog Agent, but can be downloaded directly from Oracle.

Note: The following runtimes are required on your system for JPype, one of the libraries used by the Agent when using JDBC Driver:

• Java 8 or higher
• Microsoft Visual C++ Runtime 2015

JDBC Driver

• Download the JDBC Driver jar file.
• Add the path to the downloaded file in your $CLASSPATH or the check configuration file under jdbc_driver_path (see the sample oracle.yaml).

Oracle Instant Client

The Oracle check requires either access to the cx_Oracle Python module, or the Oracle JDBC Driver:

1. Go to the download page and install both the Instant Client Basic and SDK packages.

   If you are using Linux, after the Instant Client libraries are installed ensure the runtime linker can find the libraries. For example, using ldconfig:

   # Put the library location in an ld configuration file.
   
   sudo sh -c "echo /usr/lib/oracle/12.2/client64/lib > \
      /etc/ld.so.conf.d/oracle-instantclient.conf"
   
   # Update the bindings.
   
   sudo ldconfig

2. Decompress those libraries in a given directory available to all users on the given machine (i.e. /opt/oracle):

   mkdir -p /opt/oracle/ && cd /opt/oracle/
   unzip /opt/oracle/instantclient-basic-linux.x64-12.1.0.2.0.zip
   unzip /opt/oracle/instantclient-sdk-linux.x64-12.1.0.2.0.zip

Datadog User creation

Create a read-only datadog user with proper access to your Oracle Database Server. Connect to your Oracle database with an administrative user (e.g. SYSDBA or SYSOPER) and run:

-- Enable Oracle Script.
ALTER SESSION SET "_ORACLE_SCRIPT"=true;

-- Create the datadog user. Replace the password placeholder with a secure password.
CREATE USER datadog IDENTIFIED BY <PASSWORD>;

-- Grant access to the datadog user.
GRANT CONNECT TO datadog;
GRANT SELECT ON GV_$PROCESS TO datadog;
GRANT SELECT ON gv_$sysmetric TO datadog;
GRANT SELECT ON sys.dba_data_files TO datadog;
GRANT SELECT ON sys.dba_tablespaces TO datadog;
GRANT SELECT ON sys.dba_tablespace_usage_metrics TO datadog;

Note: If you’re using Oracle 11g, there’s no need to run the following line:

ALTER SESSION SET "_ORACLE_SCRIPT"=true;

Configuration

Host

Follow the instructions below to configure this check for an Agent running on a host. For containerized environments, see the Containerized section.

1. Edit the oracle.d/conf.yaml file, in the conf.d/ folder at the root of your Agent’s configuration directory. Update the server and port to set the masters to monitor. See the sample oracle.d/conf.yaml for all available configuration options.

   init_config:
   
   instances:
    ## @param server - string - required
    ## The IP address or hostname of the Oracle Database Server.
    #
    - server: localhost:1521
   
      ## @param service_name - string - required
      ## The Oracle Database service name. To view the services available on your server,
      ## run the following query:
      ## `SELECT value FROM v$parameter WHERE name='service_names'`
      #
      service_name: "<SERVICE_NAME>"
   
      ## @param user - string - required
      ## The username for the user account.
      #
      user: datadog
   
      ## @param password - string - required
      ## The password for the user account.
      #
      password: "<PASSWORD>"

2. Restart the Agent.

Only custom queries

To skip default metric checks for an instance and only run custom queries with an existing metrics gathering user, insert the tag only_custom_queries with a value of true. This allows a configured instance of the Oracle integration to skip the system, process, and tablespace metrics from running, and allows custom queries to be run without having the permissions described in the Datadog user creation section. If this configuration entry is omitted, the user you specify is required to have those table permissions to run a custom query.

init_config:

instances:
  ## @param server - string - required
  ## The IP address or hostname of the Oracle Database Server.
  #
  - server: localhost:1521

    ## @param service_name - string - required
    ## The Oracle Database service name. To view the services available on your server,
    ## run the following query:
    ## `SELECT value FROM v$parameter WHERE name='service_names'`
    #
    service_name: "<SERVICE_NAME>"

    ## @param user - string - required
    ## The username for the user account.
    #
    user: <USER>

    ## @param password - string - required
    ## The password for the user account.
    #
    password: "<PASSWORD>"

    ## @param only_custom_queries - string - optional
    ## Set this parameter to any value if you want to only run custom
    ## queries for this instance.
    #
    only_custom_queries: true

Containerized

For containerized environments, see the Autodiscovery Integration Templates for guidance on applying the parameters below.

Validation

Run the Agent’s status subcommand and look for oracle under the Checks section.

Custom Query

Providing custom queries is also supported. Each query must have 3 parameters:

Optionally use the tags parameter to apply a list of tags to each metric collected.

The following:

self.gauge('oracle.custom_query.metric1', value, tags=['tester:oracle', 'tag1:value'])
self.count('oracle.custom_query.metric2', value, tags=['tester:oracle', 'tag1:value'])

is what the following example configuration would become:

- metric_prefix: oracle.custom_query
  query: | # Use the pipe if you require a multi-line script.
    SELECT columns
    FROM tester.test_table
    WHERE conditions
  columns:
    # Put this for any column you wish to skip:
    - {}
    - name: metric1
      type: gauge
    - name: tag1
      type: tag
    - name: metric2
      type: count
  tags:
    - tester:oracle

See the sample oracle.d/conf.yaml for all available configuration options.

Data Collected

Metrics

Events

The Oracle Database check does not include any events.

Service Checks

oracle.can_connect Verifies the database is available and accepting connections.

Troubleshooting

Need help? Contact Datadog support.