Performance insights mandatory for RDS instances to configure Integrated alerting?

Hi - I’m trying to configure integrated alerting for my MySQL RDS instances and when I add them , I do not see all the details being gathered for some of my instances even though they have enhanced monitoring enabled on them . They do not have performance insights enabled . Is that needed as well in order to get alerts and gather metrics ? Please suggest anyone ?

others that Im able to see the metrics and get alerted on are also added the same way except the only difference is they have performance insights enabled on them.

Hi Deepthi,

It looks like the enhanced monitoring hasn’t been enabled according to provided screenshot.
Could you check amount of collected unique metrics for one monitored RDS instance?
It can be done on Explore page.

count(count by (__name__) ({__name__!~"mysql.+|go.+|scrape.+",node_name="your_RDS_node_name"}))

Here is a result for an instances with disabled enhanced monitoring.

So it’s fetched only 26 node related metrics.

But you may test alerting with build-in template “MySQL connections in use”. It uses mysql metrics.

1 Like

Hi adivinho - Here is screenshot using the above query . It looks like it fetched 42 metrics

and I’m sure enhanced monitoring is enabled. I’ve verified in my AWS console. Any other suggestions on might be wrong ? mysql connections high alert seems to be working as expected ,so it an issue only with OS related alerts . Also is the built in node high memory usage alerts working ?I remember it had issues in the past with the usage of “lt” symbol . I’m using PMM 2.15.1 version as of now

1 Like

I know what you mean . Its fetching about 142 metrics for hosts that are fetching all of the data . Sooooo, what could I be missing ? It does have enhanced monitoring enabled.

1 Like

How did you add your RDS instances to PMM? Via the UI or the API? There’s a bug in the API docs that steps over the JSON needed to enable node level stats which would also give all the N/A’s although seeing CPU for your RDS instance is odd since I think both CPU (first column) and Memory (second column) come from the node…so curious why one is populating and the other is not?

1 Like

Has some service still shown empty elements now?
Maybe it was hit some month quotas for metrics at the month end.

The issue has already fixed.

1 Like

Hey advinho -

Has some service still shown empty elements now?

Yes, some of the hosts as seen above are still showing empty. This is true for one particular region (AZ in AWS ) . Other AZ’s are showing up fine .

Maybe it was hit some month quotas for metrics at the month end.

I don’t understand this . Are you talking about the quota on the AWS side ?

The issue has already fixed.

Oh ok cool ! I asked because I saw alerts from all other built-in templates but not the node high memory. Is there a separate template to be used for RDS hosts , like how you share the CPU template with me earlier ?if yes, can you please share it with me

Also , when I can expect the IA to provide better readability for the alerting ? I understand its still in Technical preview state , but just checking if you have a timeline on when this can be GA .Also any plans on adding more builtin templates or would this be all even in the GA version ?

1 Like

Could you inform us what a region is used for the instance with missed metrics?
I will try to reproduce the issue on our side.

1 Like

Thanks much for looking into it . The Virginia region is the one I’m not seeing metrics for . Oregon and Ohio regions are doing fine . If you cannot reproduce , then its probably something on my end

1 Like

I have checked an RDS on us-east-1 region and it manages to generate 121 unique metrics.

One question crossed my mind - has you added all instances by Amazon RDS access keys that are belong to the same user?
User could have permissions that are not allow to collect enhanced metrics.

1 Like

hmm … I’m using the same user/RDS access keys for all the hosts, but that thought crossed my mind as well. Could there be any other settings that could be blocking from collecting the metrics ? other than the user

1 Like

Please check used policies. Here you may find how-to create/check policies:
https://www.percona.com/doc/percona-monitoring-and-management/1.x/amazon-rds.html#creating-a-policy

1 Like

hey deepthip …what is the solution…please explain

1 Like

how you solved this problem?

1 Like