Step-By-Step Create OEM Metric Extension - Send Backup Failure Alerts Using Metric Extension

[@xxkjudth]

Hello, first of all, thank you very much for the video, it has been very helpful, how could I implement this so that it also reports when a backup was successfully executed?

[@YouVolve]

Hi, thanks for your question. Yes, you can generate an alert when a backup was successfully executed. Please take the last SQL query in the shared doc in my Google drive (link in the description of the video). That query returns three different values based on the backup job status: SUCCESS, WARNING and FAILED. In your case take the value SUCCESS instead of the other two in the Metric Extension's variable field. You will receive the SUCCESS status reported by the ME.

[@xxkjudth]

@@YouVolve Hello, how can I modify the alerts that arrive by email? because when you put success it comes with an event warning

[@YouVolve]

Hi Karen, the ME in the tutorial is designed only for handling backup failures. So, it has only two alerting options: Warning or Critical. If you need a notification for success status, you have to create a separate ME with a different category like informational and write the appropriate SQL query for that.

[@erics5944]

Hi Manash, great script! I will try to adapt it for our Data Guard environment (backups are take from the primary databases) so I can add the ME to the instance monitoring template and won't get criticals from the standby databases.

[@YouVolve]

Yes, definitely and monitoring template is the best way to implement it.

[@krishnamoorthiece8558]

Can we create metric for database full backup as well?

[@YouVolve]

Thanks Krishnamoorthi for your question. Metric or Metric Extensions are nothing but some statistics which can be used to trigger alerts based on what value they have at a certain point in time. It cannot be used to perform a full backup or an activity like that. If you want to perform an activity like a full backup then you have to create a job. I have a 3 part video on how to automate RMAN backups and reporting in OEM using OEM jobs. If you are interested then you can watch them with the links below: Part-1 : kzbin.info/www/bejne/aZWWlaxjZb1jb6s Part-2 : kzbin.info/www/bejne/ooPIgYqtqr2fqJI Part-3: kzbin.info/www/bejne/mHOvl3x3fseqgM0

[@krishnamoorthiece8558]

​@@YouVolve Thanks for the reply...I don't want to perform full backup.I just need to trigger the full backup status alert same as you created metric extension for Archive backups.

[@YouVolve]

Yes, you can create the alert for full backup status also. Just change the filter condition in the WHERE clause of the SQL used in the metric extension from "INPUT_TYPE='ARCHIVELOG'" to "INPUT_TYPE like 'DB_%'" so it will include both incremental and full database backups for the alerts.

[@krishnamoorthiece8558]

​@@YouVolve Thanks for the clarification ☺️

[@ccxu2926]

Hi，Alert Threshold： Warning can be configured to contain multiple values. Will an alarm be generated if one of these values is matched?

[@YouVolve]

Hi CC, if I understood your question correctly, you want to configure multiple values as Warning threshold and send the alert if at least one of the values matches right? But, here you cannot use multiple values. Instead, you can use pattern or regular expression if needed if you are comparing with strings.

[@alexphatelevich3876]

Hi Thank you very mach for this video!

[@YouVolve]

Thanks for your feedback Alex.

[@anthoniangha6539]

Great Video. I however did not see the metrics under my target after deploying it.

[@YouVolve]

Thanks...a deployed ME won't appear in the metric page if the target type you are searching in is different than the deployed target type. For example, if your ME is for DB instance, it won't appear at DB cluster level. But if this is not the case, then probably you should raise a case with Oracle support as there is no way a deployed ME won't appear in the metrics collection and settings options for a target.

[@edwincristianoviedobertel1419]

What is the difference between creating a Set Rule of type Enterprise and Personal? How do we set alerting priority?

[@YouVolve]

Hello Edwin, Enterprise rule sets are enforced at enterprise level that includes creating/updating incidents, and sending alerts to multiple users/admins. Personal Notification rule sets are useful for admins to be notified about changes the admin is interested in. They can only be used to send alert to the owner of the rule set. Priorities are defined at individual rule level and severity is assigned by OEM itself. You can utilize the severity to determine what action to take against each of the rule you create under a rule set. You will get options for both while creating a rule set.

[@christianoviedo6025]

@@YouVolve Is there any chance that a rule will step on another rule, since I made a personal rule, but did not alert my email

[@YouVolve]

Hello Christian, A rule will not step on another rule. They will just execute as per the order they were created within the ruleset. For example: it is possible that two rules under the same rule set can take a duplicate action like sending email to the same recipient. In that case the receiver will get two emails for the same incident. Whether the scope is Enterprise or Personal it is defined at ruleset level not in the rule level. So you cannot a create a ruleset with both enterprise or personal rules within it. However, you can create two rulesets - one as enterprise and another as personal. Both will execute without interfering each other.

[@christianoviedo6025]

@@YouVolve Thank you very much for your help, it was really urgent.

[@alexphatelevich3876]

In --SQL used in Metric Extension to extract archivelog backup status and detect if backup did not run in the last 2 hours you missed 3 condition : WHEN STATUS = 'RUNNING' THEN 'RUNNING' Otherwhise it will throw : CRITICAL

[@YouVolve]

Let me take a look and get back today...

[@YouVolve]

Thanks Alex for pointing out the flaw. I have introduced a new condition in the WHERE clause as "STATUS NOT LIKE '%RUNNING%' " to keep the running backups out of the context as adding the "WHEN STATUS = 'RUNNING' THEN 'RUNNING'" will cause the ME to send a false "Clear" alert if the previous run of the job failed and as we never know if a running backup will end successfully or not. You can download the updated SQL if you want and please let me know if you face any issues.

[@shymonr1009]

can you create a video for entire database monitoring --incident rules and email notification

[@YouVolve]

Hi Shymon, thanks for your interest. I am working on a full course on OEM that covers starting with preparing the host, installation, monitoring setup, incident rules, automations, upgrade, patching etc, everything a DBA needs to know to be an expert. The course will have two parts covering basics to the advanced topics and will be published in Udemy during March/April 2022. Unfortunately Udemy courses cannot be published in KZbin as it is a free platform. However, I will continue to publish videos in KZbin covering individual topics that is useful for DBAs.

[@edgardonb]

Dude,,,,, how do you create the rule? I don't see where you relate it

[@YouVolve]

Hello Edgard, creating an incident ruleset is altogether a different topic which I am going to cover in another video. If you are using OEM, you already have at least one ruleset configured.

[@oekpet4768]

Hello, how to send deadlock ?

[@YouVolve]

Hi Oek, thanks for your interest. I am working on uploading a video on that within next 2-3 weeks...

[@oekpet4768]

@@YouVolve big thank

[@anthoniangha6539]

I created a metrics extension using the V$RMAN_BACKUP_JOB_DETAILS and deployed it to all targets. I however is looking at the option to use this metrics to generate a daily backup report using the mgmt$metric_current view following Doc ID 1622443.1 and 2853134.1 Do you have a script for this?

[@YouVolve]

Hi Anthonia, I don't have a script ready for your purpose but you can use the SQL below: select target_name,value from sysman.mgmt$metric_current where metric_name like '' and column_label='' Please include other column in the SELECT list where you think useful. Hope this will help.