Veeam Reporter

[ivordillen]

hello,

I saw last weekend that I had some high IO intensive tasks (a lot of IOPS write) and that has an impact on my infrastructure (HP EVA Raid5).

I had a very good view in veeam monitor but when I search for the same period now (3 days later) there is much less details. I also see too little details in Reporter where I thought I was told (months ago) that I should find such detailed older data.

When I need to troubleshoot or when I want to look for bottlenecks I need detailed data to find the raison for high IO latency (example)

kind regards,

Ivor

[Vitaliy S.]

Hi Ivor,

Veeam ONE aggregates the longer term performance to keep the size of the database manageable. Our default configuration of the data stamps is the following:

- 20 seconds for the past hour
- 5 minutes for the past week
- 2 hours data stamps for everything more than 1 week

These values are the most optimal for troubleshooting and database management purposes. Extending these aggregation intervals will lead to huge database growth and will cause much more load on your existing SQL Server.

Thanks!

[ivordillen]

I am really wondering if I am the first or only person who would like to have more details for a longer period...

For instance at night we have a lot of backups (veeam, ibm TSM, SQL dumps,...) this gives a lot of strain on my storage devices. So I want to examine the nightly periodes (21:00 - 06:00) deeper to see where the datastore latency peaks, where there are vm's that are doing very high read/write IOPS.

A data stamp of 5 minutes gives me to much topped of spikes.

For some this could be enough, for some it should be more adaptive to change like in the vcenter. When you have an issue you can alter it towards more and longer details and when everything is gooing smooth you can change it back.

Like i see it all the data is entered the database but cleaned up. If it is just a database growing issue I think we can manage. When performance drops this is someting else.

kind regards

[Vitaliy S.]

You're not the first one to ask about more granular data. Based on the Monitor 5 feedback we've changed our default aggregation intervals in Veeam ONE v6, so right now we are displaying data in a more granular way than vCenter Server can currently show.

Btw, could you please tell me which counters should be collected in a more granular way (in your opinion)? What aggregation intervals would you like to have? Also what is the size of your infrastructure? How many hosts/VMs/datastores do you have?

Thanks!

[ivordillen]

Vitaliy,

I don't think there is a quick reply but I suppose I would like to be able to have 24h full detailed data to work with.

My virtual infrastructure exists of 10 esx hosts, 240 vm's, 53 datastores with a total of 22TB

kind regards,

Ivor

[Vitaliy S.]

Ok, thanks for the feedback.

[Daveyd]

+1 in regards to Ivor's suggestion. A full 24 hours worth of granular data would be very useful.

[jeremyh8]

I would like to get more granular data also. how about giving us the option to tune the intervals? It could even be granular enough so we could set the statistic level per object. some vms are more important than others.

[Vitaliy S.]

Jeremy, tuning intervals might lead to significant growth of the database. Could you please tell me if you're using a remote full blown SQL Server to host Veeam ONE database? What intervals would you like to have by default or you haven't thought about it yet?

[jeremyh8]

we use a remote full blown sql database. I would look at 20 seconds for 24 hours then 5 minutes for 30 days then 2 hours for 1 yr and daily for year 2. I think it would be nice to set these options per buisness view group so that high priority machines can be more granular longer. This would help keep database growth down.

[Vitaliy S.]

Ok, got it. Thanks for the feedback!