Return
-
Title
How do the Usability Connection Availability and Usability OS Connection Availability connect to databases? -
Description
Usability Connection Availability
- Does Foglight use telnet to probe the database availability?
- If not, does Foglight really connect to the database instances and select something from a system database, and then if it's connected, the connection will be dropped completely before establishing next connection? How often?
Usability OS Connection Availability
- Does Foglight use ping to see whether the server is available? If not, what is it?
- How frequently does Foglight trigger the probing?
-
Cause
The connection availability rules are used to check instance connectivity, rather than whether the instance is up or down -
Resolution
Usability Connection Availability
- This metric only checks if a new (not an open one in the agent connection pool) JDBC connection can be established to the monitored instance.
- No SQL is in used for that metric in both Oracle and SQL Server cartridges.
- The SQL that is being executed is used for the 'respond' metric (check the time it takes to run a simple SQL)
For Oracle
Alarm Instance Status Trigger Alarm Description Alarm Detail DBO - Usability Connection Availability mounted or down true This alarm fires when the connection to the instance has failed. In a single instance database environment, users will not be able to connect or retrieve any information from the database. The connection to the instance INSTANCENAME failed. Failed to establish a Oracle JDBC connection protocol to Host [AGENTNAME],
user account [system] . Reason : [Listener refused the connection with the following error:
ORA-12528, TNS:listener: all appropriate instances are blocking new connections -
Profile:OracleProfile{host='HOSTNAME', service='SERVICENAME', username='system', asSysDBA=false, ports='1521', properties={oracle.jdbc.ReadTimeout=900000}}]. . [Oracle Error Code = 12528. SQL state : 66000].DBO - Usability Availability down true This alarm fires when the percentage of nodes available in the cluster falls below a predefined threshold. When one or more nodes in the cluster are not running (down), the cluster load balance is affected, thereby decreasing the database's overall survivability. 0.00% of the Cluster nodes are available DBO - Usability Availability Single down true This alarm fires when the database instance is not running (down). In a single instance database environment, users will not be able to connect or retrieve any information from the database. The Instance INSTANCENAME is down. When using a normal Oracle user and the database status is mounted, the JDBC connection will not work , the result should be an alarm that the "database is down".
When using a Sysdba user and database status is mounted, JDBC should work fine so an alarm will not be triggered (however jdbc_connect_availability_raise_alarm will set to false to in the agent log).Usability OS Connection Availability
- This metric checks both connection establishment and query execution.
- For Oracle monitored instance on Windows, it execute the following WMI query (the '0' is being replaces with the instance name): select state from Win32_Service where Name = ''OracleService{0}''
- For MSSQL monitored instance, it execute the following WMI query (the '0' is being replaces with the instance template name) select state from Win32_Service where Name = ''{0}''
- For Oracle monitored instance on UNIX, it execute the following query (the 'instance_name' is being replaces with the instance name):
ps -ef | grep -iw ora_pmon_$instance_name | grep -vc grep