Advertisement
Not a member of Pastebin yet?
Sign Up,
it unlocks many cool features!
- Ashley Kuritz
- ,
- 09:15
- ,
- @Scott Coleman @Bill Riemers
- Kelly McGeedy has brought up the latancy issue, and Chandra wants us to look at it closer with her, to see if we can find a solution
- She has described the issue happening on 3 different opportunities, at the same area.
- I can't give any more details, because that is the extent of what I have been given
- I was hoping we could jump on a call with Kelly and have her simulate the issue.
- Scott Coleman
- ,
- 09:21
- ,
- My first availability for a real-time call is 2pm but I know this is important - could Kelly record a Bluejeans video that shows her going through the steps where she has performance issues and share it with us?
- Ashley Kuritz
- ,
- 09:22
- ,
- good Idea, let me talk to her
- Scott Coleman
- ,
- 09:22
- ,
- Edited,
- Then I could review this morning possibly between meetings
- Just rechecked my calendar and I am now open at 10am today for a live call
- Bill Riemers
- ,
- 09:37
- ,
- Kiran has a configuration update to reduced the latency by about 40%. That is really the only "quick" fix. But we have lots of less dramatic improvements we could do that could combine to have a big effect.
- We don't really need the problem simulated again...
- Ashley Kuritz
- ,
- 09:38
- ,
- Good news! okay, however, Kelly has thrown this up to Essam who wants us to look at her specific issue, the configuration may help, but to appease them can I still have her record it, or have you seen her specific issue already?
- Scott Coleman
- ,
- 09:40
- ,
- Please have her record it - I don't want to assume we know what this issue is she's seeing
- Ashley Kuritz
- ,
- 09:40
- ,
- okay, thanks 🙂
- Bill Riemers
- ,
- 09:42
- ,
- That seems reasonable to me.
- Ashley Kuritz
- ,
- 09:54
- ,
- They are asking that you guys see it live as well, I added you to the invite, however she is working on the demo now (recording it)
- if I can just come to the table with we hear you, there's no need to do this live, and this is what we propose, I think that may be a better use of everyone's time.
- **Correct me if i am wrong**
- Scott Coleman
- ,
- 10:05
- ,
- I don't see any additional value in seeing it live - I would just be there to capture a recording so I can share the example with anyone else engaged on the performance issues.
- I appreciate Kelly stepping up and being willing to reproduce the issue for us
- Bill Riemers
- ,
- 10:07
- ,
- The main value of seeing it live would be if it turns out it isn't what we are expecting, and it isn't an apple to apple comparison, maybe we could suggest better things to look at. But in that case, I would personally prefer the chance to review and really think through my suggestions first.
- Ashley Kuritz
- ,
- 10:10
- ,
- okay, i hear both of you 🙂 stand by, and as soon as I get it- I will send it your way. You both should have the invite for the 1:30 call, if you can make it that would be great.
- Bill Riemers
- ,
- 55 mins
- ,
- I just forwarded you the both the previous discussions around, what I think is the same case.
- Ashley Kuritz
- ,
- 50 mins
- ,
- okay, just sent the recordings your way
- check your email
- Scott Coleman
- ,
- 50 mins
- ,
- Thanks Ashley!
- Ashley Kuritz
- ,
- 49 mins
- ,
- You're welcome!
- Scott Coleman
- ,
- 32 mins
- ,
- I will continue to say this to whoever will listen - I don't think these are issues we will solve realtime in meetings with users - I think we need to staff a performance initiative to really dive into this
- So I wouldn't expect to have any quick fixes at 1:30
- Bill Riemers
- ,
- 29 mins
- ,
- Just looked. Indeed this is the same as the case I forwarded. Only this is the renewal scenario. Our slowest case for renewals is many public sector, many lines, where incentives apply...
- Bill Riemers
- ,
- 26 mins
- ,
- So with batch size update we would drop this from 3:20 to about 2:10. A second think we can do for this worst case scenario is rewrite the target discount flow. But that only impacts the very initial cart load. It could potentially reduce that time by an additional 40%.
- Ashley Kuritz
- ,
- 25 mins
- ,
- okay, the batch size update-- is that what Kiran has created? or is that somethihng different
- Bill Riemers
- ,
- 24 mins
- ,
- Yes. Kiran has the batch size update ready, and that is what we tested in our e-mail. We really just need the green light to activate that in production, as it is a single configuration variable update.
- Ashley Kuritz
- ,
- 24 mins
- ,
- understood
- that still doesn't really get us where we need, it's improvement. But Chandra is wanting this done in seconds, not minutes.
- Bill Riemers
- ,
- 22 mins
- ,
- The flow update we haven't worked on, because it would need to be a story. Basically what happens today is when we apply the target discount for a renewal incentive, we reprice again. In theory we have all the data we need just to calculate the prices. Only we didn't do it that way because we felt more confident allowing Apttus to do the calculations. But it might speed up the initial cart load significantly to take the more complicated approach of recalculating the prices without relying on Apttus to do it.
- Ashley Kuritz
- ,
- 21 mins
- ,
- I agree with Scott's approach that with this performance issue we need to look at potentially creating a "mini" project team to look at this, and create + Complete.
- Scott Coleman
- ,
- 20 mins
- ,
- Devote some people to this full time, free them of their other responsibilities, these are actions that would make this a clear top priority. Right now it seems like we're trying to rely on people's spare cycles to solve this. I don't think that will be successful
- Bill Riemers
- ,
- 20 mins
- ,
- Yeah. That is the problem. It is impossible for any of us to work on this without dropping something else on the floor...
- Ashley Kuritz
- ,
- 20 mins
- ,
- okay
- let me bring that up to Chandra
- Scott Coleman
- ,
- 20 mins
- ,
- Bold reprioritization is needed
- Ashley Kuritz
- ,
- 19 mins
- ,
- indeed. We will never get this right unless we pivot.
- Bill Riemers
- ,
- 19 mins
- ,
- We have plenty of data to identify issues. For example, everytime I price a scenario like these I see the following in my e-mail:
- ApexApplication <info@salesforce.com>
- Mon, Jun 24, 3:22 PM (19 hours ago)
- to briemers+stage@redhat.com
- Operation: /apexremote
- By user/organization: 00560000000mStZ/00D190000009U62
- Caused the following Apex resource warnings:
- Apttus_Config2:Number of query rows: 33030 out of 50000
- (these emails can be disabled from the user detail page for this user)
- Operation: /apexremote
- By user/organization: 00560000000mStZ/00D190000009U62
- Caused the following Apex resource warnings:
- Apttus_Config2:Number of SOQL queries: 54 out of 100
- (these emails can be disabled from the user detail page for this user)
- These type of messages tell us we have way too much processing going on while pricing...
- But it takes a lot of analysis and hard work to determine from that what we can actually work on.
- Scott Coleman
- ,
- 18 mins
- ,
- I can join the call with Kelly and Essam at 1:45
- Ashley Kuritz
- ,
- 18 mins
- ,
- okay, good.
- I see what you are saying Bill
- For a team what resources would we need?
- Bill Riemers
- ,
- 17 mins
- ,
- But we also need to set realistic scenarios. If you have dozens of lines in the cart, we probably won't get the time below 1 minute, even if we strip out all our custom logic.
- Ashley Kuritz
- ,
- 17 mins
- ,
- 1 tech, 1 BSA?
- Bill Riemers
- ,
- 16 mins
- ,
- tech, BSA, QASource, Product Owner. I think that is a minimum team.
Advertisement
Add Comment
Please, Sign In to add comment
Advertisement