This agent can monitor several parameters from an Oracle database instance. We've tested it against Oracle release 8, 8i, 9i and 10g, but we strongly recommend you to test it before implementing the final monitorization in production environments.
The Oracle agent has been developed using the functionalities and improves of the Osmius framework and ACE libraries, so it's mandatory the installation of ACE libraries for the correct Osmius installation and execution. This agent also needs Oracle libraries to use the API's known as Oracle C++ Call Interface, you need OCCI libraries and/or Oracle “normal” or “instant” client.
ORACL001 agent provides about a dozen of events with customizable parameters to grow up and expand in a very easy way. The events has been selected according to the experiences and suggestions of the best Oracle DBA group who we know, avoiding no-sense variable monitoring.
The Oracle Osmius agent is prepared to reuse database connections and only try to reconnect when one of the scheduled events fails.This agent is made by cross-platform code and is non-intrusive. You don't need to install the agent in the same machine where the Oracle instance is. The Oracle Osmius agent is prepared to reuse database connections and only try to reconnect when one of the scheduled events fails.
Note to limit the consumption of resources of the agent in an Oracle instance.
Oracle API doesn't provide a way to implement a good “timeout” for locking operations against the database. This situation could happen for example when a database with problems is being mounted and dismounted or by excessive load.
One way to limit the excessive time of resources use in Oracle is using “oracle user profiles”:
CREATE PROFILE LIMIT_OSMIUS_QUERY_TIME LIMIT CPU_PER_CALL 1000; ALTER USER OSMIUS_USER PROFILE LIMIT_OSMIUS_QUERY_TIME;
Note To configure the Oracle libraries in Osmius.
You must specify the path of the Oracle libraries with LD_LIBRARY_PATH variable in order to allow Osmius work properly . This requires modifying the file osm.sh located in the path $OSM_ROOT\osmius\bin\osm.sh, for example:
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:$OSM_ROOT/lib:/opt/app/oracle/db/lib
Generally speaking each Osmius agent is able to monitor many instances but only one instance type. If you need to review this concept take a look at the Concepts Glosary. Each instance is individually defined in the agent configuration file (more in Agents & Instances, and depending on the Instance Type you must define certain connectio info.
The connection_info is required information to establish connections with Oracle instances. (More at connection_info)
In this case the connection_info format in this:
CONNECTION_INFO= -u <USER> -p <PASSWORD> [-d <ORACLE_SID>]
Where:
Sample:
CONNECTION_INFO= -u osmius -p pass3344 -d ORACLE_TEST
Here you have a brief description of this agent capabilities. Each event is explained in detail in next sections.
| EVENT | DESCRIPTION | c | w | a | tseconds | Extra Parameters / Comments |
|---|---|---|---|---|---|---|
| AVAILABL | Connection to instance and SELECT to DUAL table or <user_table>. | 1 | 0 | 0 | 60 | -U user_table Silent mode ( -s) suggested |
| ORSELCNT | Number of registries in DUAL Table or <user_table>. | 0 | 10 | 20 | 60 | -U user_table |
| ORNUMUSU | Number of user sessions connected to the instance. | 0 | 30 | 50 | 300 | Interesting to track evolution. |
| ORPRCCUR | Used percentage of cursors in highest consumption session. | 0 | 90 | 95 | 300 | Silent mode ( -s) suggested |
| ORNUMLCK | Number of locking global sessions. | 0 | 2 | 5 | 60 | Interesting to track evolution. |
| OROBJINV | Number of invalid objects. | 0 | 10 | 100 | 3600 | Silent mode ( -s) suggested |
| ORCHATBL | Tables with chained rows and possible performance problems. | 0 | 1 | 2 | 86400 | Silent mode ( -s) suggested Interesting: Analyze stats |
| ORBUFRAT | Hit ratio in data cache. | 1 | 98 | 95 | 60 | Interesting to track evolution. |
| ORPARRAT | Hit ratio in instructions and query parsing cache. | 1 | 90 | 85 | 600 | Interesting to track evolution. |
| ORPRCTBS | Tablespace % occupation percentage. | 0 | 95 | 98 | 1200 | -X 'TBS1'[,'TBSn] Silent mode ( -s) suggested |
| ORMAXEXT | Number of remaininf extents to reach the maximum allowed. | 1 | 55 | 2 | 1200 | Silent mode ( -s) suggested |
| ORNXTEXT | Object with no space left to increase its size. | 0 | 1 | 1 | 600 | Silent mode ( -s) suggested |
| ORPRCPRO | Used percentage of maximum number of allowed processes | 0 | 85 | 98 | 300 | Interesting to track evolution. |
| ORREDLOG | Number of redo logs not yet archived | 0 | 3 | 5 | 60 | Interesting to track evolution. Use only in databases in ARCHIVE LOG mode. |
| ORSPCDAT | Data segments space used in MBytes. | 0 | 99999 | 99999 | 86400 | Interesting to track evolution. |
| ORSPCIDX | Index segments space used in MBytes. | 0 | 99999 | 99999 | 86400 | Interesting to track evolution. |
| ORLONOPS | Long operations. | 0 | 3600 | 3600 | 86400 | Interesting to track evolution. |
| ORARAUPC | Archive Area Used (%) | 0 | 80 | 90 | 900 | -L <archive logs directory> |
| ORDATAFL | Data Failure Detected | 0 | 1 | 1 | 300 | |
| ORBRKJOB | Broken Job Count | 0 | 1 | 1 | 300 | |
| ORPRCLMT | Process Limit Usage (%) | 0 | 80 | 90 | 60 | |
| ORSESLMT | Session Limit Usage (%) | 0 | 90 | 97 | 60 | |
| ORDPAUPC | Dump Area Used (%) | 0 | 95 | 97 | 900 | -L <dump area directory> |
| ORFRAUPC | Usable Flash Recovery Area (%) | 1 | 15 | 3 | 900 | |
| ORINSTAT | Instance Status | 0 | 1 | 1 | 60 | |
| ORDBLCOR | Data Block Corruption | 0 | 1 | 1 | 300 | |
| OROUTMEM | Out of Memory | 0 | 1 | 1 | 300 | |
| ORRDLOGC | Redo Log Corruption | 0 | 1 | 1 | 300 | |
| ORARCHUN | Archiver Hung | 0 | 1 | 1 | 300 | |
| ORFRARPC | Recovery Area Free Space (%) | 1 | 15 | 3 | 900 | |
| ORTBLS00 | Tablespace Used (%) | 0 | 85 | 97 | 300 | -L <tablespace name> |
Checks the connection availability against the database instance and performs a “SELECT” sentence over a table and recovers the result.
Extra parameters:
You can use an optional parameter:
| PARAMETER | MEANING | Mandatory? |
| -U | Target table of “SELECT” sentence. Table must exist and user will need read permissions. Table must have at least one row. If the parameter is not defined this event will use the DUAL system table. | No |
Returned values:
| VALUE | Description |
| 1 | OK |
| 0 | Not available |
| -1 | Error |
Recommended Parameters:
| Comparing | Direct. The higher value the more critical the event. (c 0). |
| Monitoring Interval | 5 secs – 1 hour –> according on how critical the instance is. |
| Warning threshold | 0 |
| High severity threshold | 0 |
Sample:
AVAILABL = -t 10 -c 1 -w 0 -a 0 -T "Oracle availability using AUTHORS_BOOK table" -U AUTHORS_BOOK
Comments: If an user table is used don't forget grant read permission to Osmius user in the Oracle database.
Performs a “SELECT COUNT” sentence over a user defined table (or DUAL) and returns the number of rows.
Extra parameters:
You can use an optional parameter:
| PARAMETER | MEANING | Mandatory? |
| -U | Target table of “SELECT COUNT” sentence. Table must exist and user will need read permissions. If the parameter is not defined this event will use the DUAL system table. | No |
Returned values:
| VALUE | Description |
| -1 | Error |
| X | Number of rows |
Recommended Parameters:
| Comparing | Direct. The higher value the more critical the event. (c 0). |
| Monitoring Interval | 60 secs – 1 hour –> according on how critical the instance is. |
| Warning threshold | X |
| High severity threshold | Y |
Sample:
ORSELCNT = -t 60 -c 0 -w 10 -a 20 -T "Rows in Test Table:" -U Test
Comments: Useful event to customize the applications monitoring focused on registries number in an y table that shows the general status of the processes. It can be used to control the pending number of registries, by example, along the week days.
If an user table is used don't forget grant read permission to Osmius user in the Oracle database.
Returns the number of user sessions connected to the instance.
Extra parameters:
No
| PARAMETER | MEANING | Mandatory? |
Returned values:
| VALUE | Description |
| -1 | Error |
| X | Number of connected users |
Recommended Parameters:
| Comparing | Direct. The higher value the more critical the event. (c 0). |
| Monitoring Interval | 300 secs – 1 hour –> according on how critical the instance is. |
| Warning threshold | X - depending of the predicted average number of users |
| High severity threshold | Y - depending of the predicted average number of users |
Sample:
ORNUMUSU = -t 300 -c 0 -w 30 -a 50 -T "The number of connected users:"
Comments: The number of users and connected sessions in a while could indicate the database activity and we can go ahead problems like lack of resources or that the sessions are not being released. This event could be used to log the number of users connected depending of the hour of day, helping to establish service level agreements.
Checks the percentage of available maximum number cursors for each session. Returns information about the sessions nearest to reach the maximum consumed cursors at the moment.
Extra parameters:
No
| PARAMETER | MEANING | Mandatory? |
Returned values:
| VALUE | Description |
| -1 | Error |
| 0 - 100 | Percentage of used cursors |
Recommended Parameters:
| Comparing | Direct. The higher value the more critical the event. (c 0). |
| Monitoring Interval | 300 secs – 1 hour –> according on how critical the instance is. |
| Warning threshold | X - depending on DBA criteria |
| High severity threshold | Y - depending on DBA criteria |
Sample:
ORPRCCUR = -t 300 -c 0 -w 80 -a 95
Comments: It's common that applications use intensively the cursors and by error or lack of experience the cursors are not being released generating errors like ORA-1000 “Maximum open cursors exceeded”. Check percentage and change schedule of cursor use or the OPEN_CURSORS parameter.
Check the number of global locks of Oracle sessions what are locking other sessions.
Extra parameters:
No
| PARAMETER | MEANING | Mandatory? |
Returned values:
| VALUE | Description |
| -1 | Error |
| N | Number of blocking locks |
Recommended Parameters:
| Comparing | Direct. The higher value the more critical the event. (c 0). |
| Monitoring Interval | 30 secs – 1 hour –> according on how critical the instance is. |
| Warning threshold | 2 - depending on activity and type of use |
| High severity threshold | 5 - depending on activity and type of use |
Sample:
ORNUMLCK = -t 60 -c 0 -w 2 -a 5 -T "The number of global blocking locks is:"
Comments: The number of locks between sessions, if the applications that use database have been well designed, this number should always be zero. In addition, if a lock appears it should be solved quickly. Session locks that persist in time can be hazardous.
Check the number of invalid objects (usually PL/SQL procedures and packages) in database and normally we need to recompile them.
Extra parameters:
No
| PARAMETER | MEANING | Mandatory? |
Returned values:
| VALUE | Description |
| -1 | Error |
| N | Number of invalid objects |
Recommended Parameters:
| Comparing | Direct. The higher value the more critical the event. (c 0). |
| Monitoring Interval | 1 hour – 1 day –> according on how critical the instance is. |
| Warning threshold | 2 - depending on activity and type of use |
| High severity threshold | 50 - depending on activity and type of use |
Sample:
OROBJINV = -t 3600 -c 0 -w 10 -a 100 -T "Invalid object are:"
Comments:
Use this sentence to compile your invalid objects:
select 'alter ' || decode(object_type, 'PACKAGE BODY', 'PACKAGE', object_type)
|| ' ' || owner || '.' || object_name || ' compile'
|| decode(object_type, 'PACKAGE BODY', ' body;', ';')
from dba_objects
where status='INVALID'
and object_type <> 'SYNONYM'
order by owner,
decode(object_type, 'PACKAGE BODY', 'AAA', 'PACKAGE', 'AAB',
substr(object_type, 1, 3)) desc,
object_name;
Invalid synonyms fall out of recompilation in this case.
Checks the number of tables with chained rows. It gives us a database fragmentation idea related to rows and performance problems.
Extra parameters:
No
| PARAMETER | MEANING | Mandatory? |
Returned values:
| VALUE | Description |
| -1 | Error |
| N | Number of tables with chained rows |
Recommended Parameters:
| Comparing | Direct. The higher value the more critical the event. (c 0). |
| Monitoring Interval | 1 day – 1 week –> according on how critical the instance is. |
| Warning threshold | 2 - depending on activity and type of use |
| High severity threshold | 10 - depending on activity and type of use |
Sample:
ORCHATBL = -t 86400 -c 0 -w 1 -a 2 -T "Number of tables with chained rows"
Comments:
Tables must be analyzed to verify if “chained rows” exist.
Our recommendation: Comment with your DBA and make an automatic analysis plan.
Percentage of Success on Data Cache
Extra parameters:
No
| PARAMETER | MEANING | Mandatory? |
Returned values:
| VALUE | Description |
| -1 | Error |
| 1 - 100 | Buffer Cache Hit Ratio |
Recommended Parameters:
| Comparing | Inverse. The higher value the lesser the severity. (c 1). |
| Monitoring Interval | 60 secs – 1 hour –> according on how critical the instance is. |
| Warning threshold | 98 - depending on activity and type of use |
| High severity threshold | 95 - depending on activity and type of use |
Sample:
ORBUFRAT = -s -t 5 -c 1 -w 98 -a 95 -T "Hit percentage in Oracle data cache:"
Comments:
You must take in consideration that the percentage will be low at the first moment of instance usage after being started up.
If buffer cache hit ratio is low it's possible for the sentences to be too open against database tables and perhaps indexes should be reviewed. Consult your DBA and/or Oracle manuals.
Percentage of Success on sentences and instructions cache.
Extra parameters:
No
| PARAMETER | MEANING | Mandatory? |
Returned values:
| VALUE | Description |
| -1 | Error |
| 1 - 100 | Parse Cache Hit Ratio |
Recommended Parameters:
| Comparing | Inverse. The higher value the lesser the severity. (c 1). |
| Monitoring Interval | 60 secs – 1 hour –> according on how critical the instance is. |
| Warning threshold | 95 - depending on activity and type of use |
| High severity threshold | 90 - depending on activity and type of use |
Sample:
ORPARRAT = -t 600 -c 1 -w 95 -a 90 -T -T "Parse Hit Ratio:"
Comments:
You must take in consideration that the percentage will be low at the first moments of instance usage after being started up.
It can give us an idea about the sentences being reused and the proper “bind” variables usage
Occupation percentage in Tablespaces
Extra parameters:
You can use an optional parameter:
| PARAMETER | MEANING | Mandatory? |
| -X | Exclude this tablespaces from the query. It accepts a tablespace name list which we are not interested in monitoring. It's usual not to monitor SYSTEM or TEMP. -X ”'SYSTEM','SYSAUX','TEMP'” | No |
Returned values:
| VALUE | Description |
| -1 | Error |
| 0 - 100 | Occupation percentage of the most occupied tablespace |
Recommended Parameters:
| Comparing | Direct. The higher value the more critical the event. (c 0). |
| Monitoring Interval | 300 secs – 1 day –> according on how critical the instance is. |
| Warning threshold | 0 |
| High severity threshold | 0 |
Sample:
ORPRCTBS = -s -t 1200 -T "Tablespaces:" -c 0 -w 95 -a 98 -X "'SYSTEM','TEMP'"
Comments:
In addition to the value this event returns the most occupied tablespace name in the event text.
It's important to predict the space problems in Oracle tablespaces. Difference cold be stopping all process and backup recover or simply give more space (Oracle is able to increase space automatically – See Oracle Administration Manual).
Object with the highest probability of reaching the maximum number of extensions.
Extra parameters:
No
| PARAMETER | MEANING | Mandatory? |
Returned values:
| VALUE | Description |
| -1 | Error |
| N | Integer value that represents remaining extensions. In the text: Owner, type and object name and tablespace where object is. |
Recommended Parameters:
| Comparing | Inverse. The higher value the lesser the severity. (c 1). |
| Monitoring Interval | 300 secs – 1 day –> according on how critical the instance is. |
| Warning threshold | 5 - depending on activity and type of use |
| High severity threshold | 2 - depending on activity and type of use |
Sample:
ORMAXEXT = -t 300 -T "Object close to maximum extensions" -c 1 -w 5 -a 2
Comments:
Objects have a MAX_EXTENTS field what limits the number of extensions that can be used according to their needs of storage.
With this event we can foresee situations where the objects that can't extend anymore, making free space for new records creating problems to processes that could be important.
Returns first object whose next extension wouldn't fit in available free space in their assigned tablespace.
Extra parameters:
No
| PARAMETER | MEANING | Mandatory? |
Returned values:
| VALUE | Description |
| -1 | Error |
| 0 | No problem. |
| 1 | Extensions problem. The text contains the owner, type, object_name and tablespace of the problematic object. |
Recommended Parameters:
| Comparing | Direct. The higher value the higher the severity. (c 0). |
| Monitoring Interval | 300 secs – 1 day –> according on how critical the instance is. |
| Warning threshold | 1 - Always detected |
| High severity threshold | 1 - Idem |
Sample:
ORNXTEXT = -s -t 3600 -T "Objects with next extensions problems" -c 0 -w 1 -a 1
Comments: When the object tries to grow up in a new extension, if this one cannot fit in the free space of the tablespace we'll get errors and stopped processes.
Percentage of used processes.
Extra parameters:
No
| PARAMETER | MEANING | Mandatory? |
Returned values:
| VALUE | Description |
| -1 | Error |
| 0 - 100 | % of total processes used in the system. |
Recommended Parameters:
| Comparing | Direct. The higher value the higher the severity. (c 0). |
| Monitoring Interval | 60 secs – 1 day –> according on how critical the instance is. |
| Warning threshold | 85 - Depending on instance type and use. |
| High severity threshold | 90 - Depending on instance type and use. |
Sample:
ORPRCPRO = -t 300 -T "Process Percentage:" -c 0 -w 85 -a 95
Comments: If this percentage is high contact your DBA to raise “maximum processes” parameter.
Returns number of unarchived “redo logs”. Activate this event only if database is in “ARCHIVE LOG” mode.
Extra parameters:
No
| PARAMETER | MEANING | Mandatory? |
Returned values:
| VALUE | Description |
| -1 | Error |
| N | Number of logs not currently archived. |
Recommended Parameters:
| Comparing | Direct. The higher value the higher the severity. (c 0). |
| Monitoring Interval | 30 secs – 1 hour –> according on how critical the instance is. |
| Warning threshold | 4 - Depending on instance type and use. |
| High severity threshold | 10 - Depending on instance type and use. |
Sample:
ORREDLOG = -t 60 -T "Number of unarchived redo logs" -c 0 -w 4 -a 10
Comments:
If redo log are not archived the all activity on database can be stopped.
Warning thresholds depends on redo logs number and size.
MB used in data segments in the whole database.
Extra parameters:
No
| PARAMETER | MEANING | Mandatory? |
Returned values:
| VALUE | Description |
| -1 | Error |
| N | Megabytes of data. |
Recommended Parameters:
| Comparing | Direct. The higher value the higher the severity. (c 0). |
| Monitoring Interval | 1 day – 1 week –> according on how critical the instance is. |
| Warning threshold | High enough not to send too much warnings. |
| High severity threshold | High enough not to send too much warnings. |
Sample:
ORSPCDAT = -t 86400 -c 0 -w 999999 -a 999999 -T "Space used in Tables"
Comments:
It could be interesting to recover this parameter once a day, week or month, to see historical evolution of used space in data base tables.
Usefull in trends and capacity planning.
MB used in index segments in the whole database.
Extra parameters:
No
| PARAMETER | MEANING | Mandatory? |
Returned values:
| VALUE | Description |
| -1 | Error |
| N | Megabytes of indexes. |
Recommended Parameters:
| Comparing | Direct. The higher value the higher the severity. (c 0). |
| Monitoring Interval | 1 day – 1 week –> according on how critical the instance is. |
| Warning threshold | High enough not to send too much warnings. |
| High severity threshold | High enough not to send too much warnings. |
Sample:
ORSPCIDX = -t 86400 -c 0 -w 999999 -a 999999 -T "MBytes of Space used in Indexes is:"
Comments:
It could be interesting to recover this parameter once a day, week or month, to see historical evolution of used space in data base indexes.
Usefull in trends and capacity planning.
Heavy sentences with high execution time.
Extra parameters:
No
| PARAMETER | MEANING | Mandatory? |
Returned values:
| VALUE | Description |
| -1 | Error |
| N | Seconds taken in the longest operation. |
Recommended Parameters:
| Comparing | Direct. The higher value the higher the severity. (c 0). |
| Monitoring Interval | 1 day – 1 week –> according on how critical the instance is. |
| Warning threshold | 3600 secs – Depends on sentences type in the instance. |
| High severity threshold | 7200 secs – Depends on sentences type in the instance. |
Sample:
ORLONOPS = -t 86400 -c 0 -w 3600 -a 7200
Comments: It could be interesting to recover this parameter once an hour or a day to identify long operations and send it to developers and dbas to fix and rewrite them.
ORARAUPC returns the Archive Area Used percentage (%).
Extra parameters:
This event need an extra parameter to work:
| PARAMETER | MEANING | Mandatory |
| -L | –L <archive logs directory> - Logs directory path | YEs |
Return values:
| VALUE | MEANING |
| -1 | Error |
| 0 - 100 | Archive Area Used percentage |
Recommended Parameters:
| Comparing | Direct. The higher value the higher the severity. (c 0). |
| Monitoring Interval | 900 secs |
| Warning threshold | 80 |
| High severity threshold | 90 |
Sample:
ORARAUPC = -s -t 900 -T "Archive Area Used (%)" -c 0 -w 80 -a 90
ORDATAFL checks Data Failure Detected.
Return values:
| VALUE | MEANING |
| -1 | Error |
| 0 | OK |
| 1 | Error Data Failure Detected |
Recommended Parameters:
| Comparing | Direct. The higher value the higher the severity. (c 0). |
| Monitoring Interval | 300 secs |
| Warning threshold | 1 |
| High severity threshold | 1 |
Sample:
ORDATAFL = -s -t 300 -T "Data Failure Detected" -c 0 -w 1 -a 1
ORBRKJOB checks Broken Job Count.
Return values:
| VALUE | MEANING |
| -1 | Error |
| 0 | OK |
| N | Number of Broken Job Count |
Recommended Parameters:
| Comparing | Direct. The higher value the higher the severity. (c 0). |
| Monitoring Interval | 300 secs |
| Warning threshold | 1 |
| High severity threshold | 1 |
Sample:
ORBRKJOB = -s -t 300 -T "Broken Job Count" -c 0 -w 1 -a 1
ORPRCLMT checks Process Limit Usage (%).
Return values:
| VALUE | MEANING |
| -1 | Error |
| 0 - 100 | Process Limit Usage (%) |
Recommended Parameters:
| Comparing | Direct. The higher value the higher the severity. (c 0). |
| Monitoring Interval | 60 secs |
| Warning threshold | 80 |
| High severity threshold | 90 |
Sample:
ORPRCLMT = -s -t 60 -T "Process Limit Usage (%)" -c 0 -w 80 -a 90
ORSESLMT checks Session Limit Usage (%).
Return values:
| VALUE | MEANING |
| -1 | Error |
| 0 - 100 | Session Limit Usage (%) |
Recommended Parameters:
| Comparing | Direct. The higher value the higher the severity. (c 0). |
| Monitoring Interval | 60 secs |
| Warning threshold | 80 |
| High severity threshold | 90 |
Sample:
ORSESLMT = -s -t 60 -T "Session Limit Usage (%)" -c 0 -w 80 -a 90
ORDPAUPC checks Dump Area Used (%).
Extra parameters:
This event need an extra parameter to work:
| PARAMETER | MEANING | Mandatory |
| -L | –L <dump area directory> - Dumps directory path | Yes |
Return values:
| VALUE | MEANING |
| -1 | Error |
| 0 - 100 | Dump Area Used (%) |
Recommended Parameters:
| Comparing | Direct. The higher value the higher the severity. (c 0). |
| Monitoring Interval | 900 secs |
| Warning threshold | 95 |
| High severity threshold | 97 |
Sample:
ORDPAUPC = -s -t 900 -T "Dump Area Used (%)" -c 0 -w 95 -a 97
ORFRAUPC checks Usable Flash Recovery Area (%).
Return values:
| VALUE | MEANING |
| -1 | Error |
| 0 - 100 | Usable Flash Recovery Area (%) |
Recommended Parameters:
| Comparison type | Inverse. The lower value the higher severity (-c 1) |
| Monitoring Interval | 900 secs |
| Warning threshold | 15 |
| High severity threshold | 37 |
Sample:
ORFRAUPC = -s -t 900 -T "Usable Flash Recovery Area (%)" -c 1 -w 15 -a 3
ORINSTAT checks Instance Status
Return values:
| VALUE | MEANING |
| -1 | Error |
| 0 | OK |
| 1 | Error |
Recommended Parameters:
| Comparing | Direct. The higher value the higher the severity. (c 0). |
| Monitoring Interval | 60 secs |
| Warning threshold | 1 |
| High severity threshold | 1 |
Sample:
ORINSTAT = -s -t 60 -T "Instance Status" -c 0 -w 1 -a 1
ORDBLCOR checks Data Block Corruption
Return values:
| VALUE | MEANING |
| -1 | Error |
| 0 | OK |
| 1 | Error |
Recommended Parameters:
| Comparing | Direct. The higher value the higher the severity. (c 0). |
| Monitoring Interval | 300 secs |
| Warning threshold | 1 |
| High severity threshold | 1 |
Sample:
ORDBLCOR = -s -t 300 -T "Data Block Corruption" -c 0 -w 1 -a 1
OROUTMEM checks Out of Memory
Return values:
| VALUE | MEANING |
| -1 | Error |
| 0 | OK |
| 1 | Error |
Recommended Parameters:
| Comparing | Direct. The higher value the higher the severity. (c 0). |
| Monitoring Interval | 300 secs |
| Warning threshold | 1 |
| High severity threshold | 1 |
Sample:
OROUTMEM = -s -t 300 -T "Out of Memory" -c 0 -w 1 -a 1
ORRDLOGC checks Redo Log Corruption
Return values:
| VALUE | MEANING |
| -1 | Error |
| 0 | OK |
| 1 | Error |
Recommended Parameters:
| Comparing | Direct. The higher value the higher the severity. (c 0). |
| Monitoring Interval | 300 secs |
| Warning threshold | 1 |
| High severity threshold | 1 |
Sample:
ORRDLOGC = -s -t 300 -T "Redo Log Corruption" -c 0 -w 1 -a 1
ORARCHUN checks Archiver Hung
Return values:
| VALUE | MEANING |
| -1 | Error |
| 0 | OK |
| 1 | Error |
Recommended Parameters:
| Comparing | Direct. The higher value the higher the severity. (c 0). |
| Monitoring Interval | 300 secs |
| Warning threshold | 1 |
| High severity threshold | 1 |
Sample:
ORARCHUN = -s -t 300 -T "Archiver Hung" -c 0 -w 1 -a 1
ORFRARPC checks Recovery Area Free Space (%)
Return values:
| VALUE | MEANING |
| -1 | Error |
| 0 - 100 | Recovery Area Free Space (%) |
Recommended Parameters:
| Comparison type | Inverse. The lower value the higher severity (-c 1) |
| Monitoring Interval | 900 secs |
| Warning threshold | 15 |
| High severity threshold | 3 |
Sample:
ORFRARPC = -s -t 900 -T "Recovery Area Free Space (%)" -c 1 -w 15 -a 3
ORTBLS00 checks Tablespace Used (%)
Extra parameters:
This event need an extra parameter to work:
| PARAMETER | MEANING | Mandatory |
| -L | –L <tablespace name> -Tablespace name | Yes |
Return values:
| VALUE | MEANING |
| -1 | Error |
| 0 - 100 | Tablespace Used (%) |
Recommended Parameters:
| Comparing | Direct. The higher value the higher the severity. (c 0). |
| Monitoring Interval | 300 secs |
| Warning threshold | 85 |
| High severity threshold | 97 |
Sample:
ORTBLS00 = -s -t 300 -T "Tablespace Used (%)" -c 0 -w 85 -a 97
The Oracle database agent has the same execution features and the other Osmius agents. See Strating and stopping agents.
| osm_ag_ORACLE01 -c osm_ag_ORACLE01.ini -m MASTERAG -p 1950 -d [>> osm_ag_oracle.log] |
Every Osmius agent can be used in standalone mode. This mode in really useful when testing a new agent in the development process.
To run in this mode yuo must add a new parameter to the configuration file (osm_ag_CODE0001.ini) named “SNDCMD”, as you can read here.
Run the Osmius agent setting the master agent communication port to 0:
| osm_ag_CODE0001 -c osm_ag_CODE00.ini -m 00000000 -p 0 -d |