Community
- Command Central
- MVP Program
- Monthly Mission
- Blogs
- Groups
- Events
- Media Vault
Products
- Observability
- Network Management
- Application Management
- IT Security
- IT Service Management
- System Management
- Database Management
Content Exchange
- SolarWinds Platform
- Server & Application Monitor
- Database Performance Analyzer
- Server Configuration Monitor
- Network Performance Monitor
- Network Configuration Manager
- SQL Sentry
- Web Help Desk
Free Tools & Trials
Store

Oracle Database & Cluster Monitoring

Hi There,

We have a requirement to monitor our Oracle database clusters and checking through Solarwinds I did find the Oracle User Experience Monitor template. As a prequistes to using i have already installed Oracle Client. However, am still left with questions as listed below;

1. Is it necessary for us to enable SNMP on OS level for Oracle monitoring...we use Redhat EL?

2. Is it necessary to configure SNMP for Oracle? If yes, can someone point me to use documents that will help me to configure this we use (Oracle 9i and 10g)

3. The oracle cluster we have also has other database. Is it possible that we can monitor only two specific db's from all that is available under the cluster? If yes, how do I specify that. I do have two tns files and they have Service Name specified under it.

4. Do we need to use the db credentials for authentication or can we use the default DBSNMP account name?

5. Do we need to add the physical node or just the listener IP to Solarwinds?

Appreciate your reply.

Thanks,

Shailesh

Find more posts tagged with

Oracle

Accepted answers

All comments

Shailesh1

Hi Moderators,

Would be grt if someone from Solarwinds could answer few of these queries :-)

Thanks

jiri.tomek

Hello Shailesh,

to answer your questions:

1 and 2. If you want to monitor Oracle from database perspective, then you need only SQL access from APM server to Oracle. APM Oracle User Experience Monitor runs SQL queries against database and gets results from these queries.

If you want to monitor Oracle also from system perspective (processes, memory consumption, ...) then you need SNMP access from APM to Oracle as well. However you don't have to expose SNMP for whole system, you can configure it to only provide Oracle data. There is some guide on http://download.oracle.com/docs/cd/E11857_01/em.111/b16244/chap2.htm

3. You specify queries used for monitoring by yourself so you can choose what to monitor. If you want to monitor only two databases, just use queries that get data for those two databases only. You also have to specify "Destination Point Name" which is used to connect to the database.

4. You have to use credentials for monitoring. These credentials are used to execute queries so they have to have appropriate permissions for data that you want to get.

5. If you don't want to monitor Oracle processes and other system data, you don't have to add Oracle computer as full node. You can use "External" node for this. But if you need to use SNMP, then you have to add it as full node.

Documentation page for Oracle User Experience monitor can be found here: http://www.solarwinds.com/NetPerfMon/SolarWinds/OrionAPMPHComponentTypesOracleQA.htm

Shailesh1

Hi Jiri,

Thanks for your reply. Can you provide me with the best practices for the thresholds for each of the Oracle component monitors as listed below:

Response Time Warning Threshold:
Response Time Critical Threshold:
Statistic Warning Threshold:
Statistic Critical Threshold:

I know its totally dependant on our enviroment....however something as a reference would be really great.

Thanks

jiri.tomek

Hello Shailesh,

Statistic threshold doesn't have any "reasonable" values in general. Statistic value is what query used for monitoring returns.

For example: If you want to monitor number of pending orders and alert if there are more than 20, you could use "SELECT COUNT(OrderID) FROM Orders WHERE IsPending = 1" and set threshold for statistic warning to 20 and critical to let's say 30.

But if you want to monitor space used by some tables, statistic results can be in thousands of bytes/kilobytes/... so thresholds would be in thousands.

Statistic threshold must be set according to what you monitor.

Response times are more "stable" but again they depend on your environment. You can try to measure average response times when database in not under heavy load and set threshold accordingly. For example average response time 100ms can be fine for a simple query but for some complicated query response time 2000ms can be average time in normal usage.

Again, you have to set thresholds according to what you monitor.

Shailesh1

Hi Jiri,

Can we monitor the below for Oracle databases from Solarwinds:

Is the database instance is up?
- Is the listener up?
- Connect as test user / system thru database is possible in a given time (e.g.1 second)
- Tablespace need to be extended by an additional datafile (whenever next extend does not fit to free space) in a next hour
- Monitoring alert-file for unexspected errors like ORA-00600 internal errors
- Backup successful executed
- Archive logs could be written successfully / space available
- Users locked by invalid username / password-usage
- review unusual behavior on buffers
- trend values e.g. load, bufferusage, etc.

Thanks