Feature Request: Properly calculate the "Duration" between events on the Job and History Session results screen.

[cvanliere]

Feature Request

Properly calculate the "Duration" between events on the Job and History Session results screen.
Highlight an event in a Job and Right Mouse (RM) click and select "Copy to Clipboard"
If you step through the each event and compare the Time Difference between the copied events. They do NOT match the listed Duration.
a) "Copy to Clipboard" each event entry and Paste these events into Excel (column D)
b) then did a =MID to extract just the Date and Time in Column C. [=MID(D58,13,8)],[=MID(D59,13,8)], [=MID(D60,13,8)], etc...
c) then compare the two in Column B with a Time Format. [=C59-C58] Fomat=30:55.2

10/28/2022 10:34:18 AM :: Preparing guest for hot backup
10/28/2022 10:44:44 AM :: Releasing guest

This shows 10 Minutes 26 Seconds between the two. The GUI shows only 14 Seconds duration. This is inaccurate and should be corrected.


We opened Case #05696680. We had a long outage (10 minutes) during a VM backup that caused our SQL Server to drop connections. Normally Veeam did not cause this kind of outage. We found out the long way (collect and send Veeam our logs) that we had an Orphaned Snapshot and Support provided "https://kb.vmware.com/s/article/2114949". Following this document we fixed our issue.

It would be nice is the system could Warn that the Duration was longer than expected. We could then research and correct on the fly.

[Mildur]

Hi Cvanliere

To be sure, you are telling Logs in %programdata%\veeam\backup and job statistics in the Veeam console doesn't not show the same duration value?
To be honest, sounds to me more like a bug than a feature request then. I cannot think of any reason why they should show different values. Let me check it and update the post if I have more information.

Thanks
Fabian

[cvanliere]

No, I am not talking about the %programdata%\veeam\backup\*.Logs, In the Veeam GUI there is a screen that shows the Results. That GUI screen has "Duration" as the last Column. In my example the Duration column says "0:00:14" (14 Seconds) after "Preparing guest for hot backup". On that exact same screen, if you Right Mouse (RM) click and select "Copy to Clipboard". Do that for the next entry "Releasing guest" and compare the two, that shows 10 Minutes 26 Seconds. I agree it is a bug, but the Ticket I opened suggested it was a Feature Request and I should post is as such. I can send screen shots and the Excel comparison if the Forum allows.

[Mildur]

Hi Cvanliere

No need to. I see the same behavior in my lab.
I believe you are talking about this duration comparison.


Thanks
Fabian

[Mildur]

Hi Cvanliere

We tested a bit internally and I talked with Support management; the timestamp you're seeing more loosely correlates to some generalized activities that the UI reports on, it's not an exact 1:1 for every action that the backup job is taking. The duration value is about how long the individual actions take as logged, but the time stamps are just referencing when the UI information appeared. There is not a relationship between the UI information and the duration value, as they're tracking different things.

So for something like "Preparing Guest for hot backup", that indeed took the amount of time for the duration. But the next event logged in the UI doesn't reflect the additional activities that happen in between that event and the previous one, it's just meant as a general guide for what the job is doing, so maths on the console time stamps are not so useful.

Ultimately, for the issue you experienced in the case, monitoring with VeeamOne is likely the best option. Parsing the logs and the UI isn't the greatest for this, and VeeamOne will capture things like orphaned snapshots, jobs running longer than usual, bigger increments than usual, etc.

Best,
Fabian

[cvanliere]

Fabian,

I still think this is a bug. If the system did 10 steps between "Preparing guest for hot backup" and "Releasing guest", then the Duration should show that, or show more steps in the UI so the duration more accurately shows the Gap we are trying to capture without the need for the full event dump. I understand the single step of "Preparing guest for hot backup" may have only taken seconds, but the true Duration between the two was Minutes in our case. The UI should show the more Accurate value.

What is VeeamOne? This is the first I have heard of it.

[cvanliere]

Fabian,

If the UI group won't budge after reading my 2:41PM post... I opened another Feature Request.
veeam-backup-replication-f2/feature-req ... 84098.html

Still interested in more Info on VeeamOne.

[Mildur]

Hi Cvanliere

Veeam One is our monitoring and reporting product.
It allows you to receive reports about your virtual infrastructure and Veeam Backup & Replication installation:
https://www.veeam.com/virtualization-ma ... ution.html

I saw that Hannes already linked to the orphaned snapshot report in the other topic.
Continue the discussion about snapshot reporting in the other topic.
Let's keep this topic about "Console Duration".

Best,
Fabian

[cvanliere]

Fabian,

Thank you for the Veeam One link. Any details you want to add?

You did not comment on my "I still think this is a bug." post.

Another idea if UI group wants to leave Duration alone. What if you ADD columns. One could be called Gap/Interval and the other Running Total

Using your example
Current UI Duration
Preparing guest for hot backup 00:07
Releasing guest 00:09


Prosed UI Duration Interval Running Total
Preparing guest for hot backup 00:07 00:17 00:17
Releasing guest 00:09 00:25 00:42

We now know the Interval between Prep and Rel was 25 seconds. Add that to the 17 seconds and the running total is 42 seconds.

[Mildur]

I will go through your answers, no worry.
Was out of office the last 2 days with no access to my lab

I'll update this topic next week.

Best,
Fabian

[cvanliere]

Fabian,

Next week is fine.

I noticed that the Post from 12/15/22 11:46AM's format was removed during the Submit. All the Blanks I put in to space the columns in the example were removed.
I also noticed you were able to insert a screen shot into the Case/Post. How do I do that?

Here is the previous Example again with Periods to align the values
Current UI............................ Duration
Preparing guest for hot backup..... 00:07
Releasing guest....................... 00:09


Prosed UI.............................. Duration..... Interval..... Running Total
Preparing guest for hot backup...... 00:07.......... 00:17........ 00:17
Releasing guest........................ 00:09.......... 00:25........ 00:42

[Mildur]

Hi Cvanliere

You can use code tags. This allows you to keep the format:

Code: Select all

Titel                    Text
One                               Text

Pictures can be added with img tags. Upload your picture to a hoster like https://imgbb.com. Use the direct link of the picture here in the forum with the img tags.



Best,
Fabian

[Mildur]

Hi Cvanliere

I discussed your request internally.
We won't add new information to the VBR console, because the additional information would be mostly helpful only to our support team. Too many actions are performed in the backend.
And if I understood it correctly, the column which you call "interval" doesn't give more value to the backup job if it only shows the interval between two timestamps from the UI. It would show the same value you can already calculate yourself between the two UI timestamps.

Best,
Fabian

[cvanliere]

Fabian,

I think I am following

Thank you for the ideas on how to get formatted responses into the thread using img tags.

Q1) If the system did 10 steps between "Preparing guest for hot backup" and "Releasing guest", then the Duration should show that, or show more steps in the UI so the duration more accurately shows the Gap we are trying to capture

A1) We won't add new information to the VBR console, because the additional information would be mostly helpful only to our support team. Too many actions are performed in the backend.

My response to A2) Understood. What if there was a Setting that could turn this Extra detail on/off. That might actually help your support team.
It could be turned on/off on the fly. When a Customer or Support Team suspects something, turn it on, refresh, and see if something jumps out as an issue. (Like our 10 minute Interval for the Orphaned Snapshot)
-------------

Q2) ADD columns. One could be called Gap/Interval and the other Running Total

A2) "the column which you call "interval" doesn't give more value to the backup job if it only shows the interval between two timestamps from the UI."
"It would show the same value you can already calculate yourself between the two UI timestamps."

My response to A2) As stated the current value in "Duration" only depicts that steps duration, NOT the interval between two console steps. Interval between the two steps timestamps (without having to manually calculate) IS what will give more value. I do not want to have to cut and paste into another application to know your true Interval. Just add it to your console.

[cvanliere]

Fabian,

Any response?

[cvanliere]

Fabian,

It has been another week with no update. What is the status?

[Mildur]

Hi cvanliere

I am very sorry. I haven't had time to go through your last reply.
I'll update this topic when I've had the time to think about it thoroughly.

Thanks,
Fabian

[cvanliere]

Fabian,

Thank you for the update. Awaiting your thoughts.

[Mildur]

Hello Cvanliere

I discussed your request again with our support management to get their input on how they use the logs and console for troubleshooting. For now our decision stays the same. We don't have plans to add new information or a toggling option to the console. It would make the console more complex with no additional benefits for analysis purpose.

Why? Our support doesn't work with the console UI when they are troubleshooting an issue. Support basically never checks the UI for timing/processing time information, that's what logs are for. And for this logs we use proper tools to analyze the content.

In general, comparing times manually is not the right way to handle monitoring like this; different application servers, different processes, all will contribute to varying times and without that domain knowledge to grant the context, this ends up just being fluff in the UI rather than useful information.
The UI should be for a quick look at statistics and details; copy/pasting anything out of it to troubleshoot infrastructure items is the wrong approach; that's why monitoring tools exist.

Best,
Fabian

[cvanliere]

I disagree, but thank you for the detail.
"The UI should be for a quick look at statistics and details", Yes, and the "Duration" is not accurate between processes shown on the screen. Inaccurate "statistics and details" should be corrected. I know you stated early on that "Duration" was for the step listed, but as I pointed out you do not show interim steps, hence inaccurate UI.

I do agree "copy/pasting anything out of it to troubleshoot infrastructure items is the wrong approach" but that is how I proved your UI was inaccurate. I do not want to cut/paste, I want the UI to be accurate. Either show all the steps so we can truly see each duration (this detail could be toggled on/off) OR if you will only show specific steps, show the duration between those OR as I suggested, show another column to differentiate.

I have said my peace. If you stand by "We don't have plans to add new information or a toggling option to the console" is you final answer, we can close this Feature Request.