Stonikbyte Great Gadgets Cycle Time and Velocity Gadgets Not responding

[JD Lobue]

Trying to load velocity and cycle time gadgets on a dashboard and it never stops processing

https://ssbinfo.atlassian.net/secure/Dashboard.jspa?selectPageId=10224

[Danut M [StonikByte]]

Hi @{5b2d16639ba6d02662e34fbd}

Can you please also provide a screenshot with the gadget settings?

Danut

‌

[JD Lobue]

‌

[Danut M [StonikByte]]

Thanks. Something seems incorrect in the settings.

You have the option “Board“ selected, but there is no board selected in the dropdown. Please select a board.

Also, you can select a filter, but make sure it is not private. Share your filter with a the project people and check if this fixes the problem.

‌

Danut.

[JD Lobue]

Still spinning on Processing

‌

‌

[JD Lobue]

Do you allow for multiple project filters in your gadgets? Here is one of the filters we are using

(project in ("SSB Central Intelligence", "SSB Client Success", "SSB Client Service Desk", "SSB Product Delivery", "SSB Discovery CRM", "SSB Business Insights CI", "SSB Core CI", "Client Success Services") AND created >= 2019-07-01 AND created <= 2020-02-09 AND "Story Points" is not EMPTY ORDER BY priority DESC)

Trying to use velocity and cycle time gadgets with the above filter hitting multiple projects is creating an ajax error, and one would think you would alert the user after minutes and not hours of the errors.

‌

[Danut M [StonikByte]]

Hi @{5b2d16639ba6d02662e34fbd}

Yes. We allow multiple projects in the filter. Please make sure that all these projects allow “Browse Project” permission for Project Role (atlassian-addons-project-access). This is needed for the app to access the issue from these projects. So please check.

Are they classic or next-gen projects projects?

We also identified some errors on our servers related to your instance. We are currently investigating.

‌

Thank you,

Danut Manda

[Danut M [StonikByte]]

Hi @{5b2d16639ba6d02662e34fbd}

‌

We did a service restart. Can you please check if the issue persists?

‌

Thank you,

Danut Manda

[JD Lobue]

I’m seeing the same ajax call error

[cid:image002.png@01D5DFF2.903A57C0]

[JD Lobue]

Great….any idea when they can be fixed? We are using both classic and nextgen, and all have browse project permissions for atlassain-addons-project-access

[JD Lobue]

I’m seeing some improvement in the response time, so I’m guessing you have done some work? However, I am still getting ajax call errors on most gadgets processing. Do we know what is causing that?

[Danut M [StonikByte]]

Hi @{5b2d16639ba6d02662e34fbd}

‌

After the restart we performed yesterday the log seems clear. Nothings seems to be wrong momentarily server side.

But from our stats it appears to be a really high usage of some the gadgets:

• Issue Formula Gadget was loaded more than 12000 times since the restart
• Cycle Time more than 9000 times
• Kanban Velocity more than 2500 times

And the usage counter seem to increase with 5 times per minute (for Issue Formula Gadget and Cycle Time Trend at least).

Is this expected? Are somehow the gadgets configured to reload automatically more times per minute? Do you perform an automatic refresh or something? If yes, I suggest increasing the refresh interval to see if it makes any difference.

‌

Thank you,

Danut Manda

‌

[JD Lobue]

We are using your gadgets quite a bit, and we have a lot of people across the org using them. We don’t have refresh on any of the gadgets, however it appears that your gadget refreshes every time the screen is loaded. What is the load performance you allow? Did you stress test your applications to allow for what we are doing? What is the maximum load?

[JD Lobue]

Can we schedule a restart after hitting a lower threshold to allow for how we are using it? We are using it as a scrolling information radiator that loads multiple times a day and cycles across a webpage an average of once per 4 minutes. Is your app able to handle this load?

[Danut M [StonikByte]]

Hi @{5b2d16639ba6d02662e34fbd}

The app and our infrastructure supports such load. But it also about your Jira instance here. Our app makes calls to Jira in order to retrieve the issue data from the filter that you specify in gadget settings. And if that filter is too heavy or it returns many issues or the requests done are too often, the Jira instance might not be able to support such a load and it becomes slow or unresponsive.

I would recommend:

• trying to increase the refresh interval of the information radiators

• optimizing the filter queries. For the filter example above you could add a few WHERE clauses that might reduce the number of issues returned:

  • “AND story_points_field is not empty” - this will exclude the issues that have no story points; they are not needed to be processed because you calculate by Story Points and the issue have have no SPs makes no difference in gadget results
  • “AND type in standardIssueTypes()“ - this will eliminate the sub-tasks from the filter result (from what I see you don't count subtasks so this would be OK)
  • “AND statusCategory = Done“ - as per your setting, the Kanban Velocity gadget counts only issues in green-color statuses (“Done“ category statuses)
  • any other queries like this that could reduce the number of issues returned by the filter

Meanwhile we will also try to optimize these gadgets by introducing some data caching (storing and then serving data from a cache once it was calculated) for reducing the calls to Jira. We already started the work on Cycle Time Trend gadget. We should have something ready soon. I will keep you updated.

Danut

‌

‌

‌

‌

‌

‌

[Danut M [StonikByte]]

Hi,

We introduced a data caching for the Cycle Time Trend Gadget. This means that once the data is calculated for a time interval it is saved in a cache and next time when the gadget is reloaded the data will be served directly from that cache. This is reducing the load on both the app and Jira and will improve the gadget’s performance. If data changes somehow (for example an issue from the gadget’s filter was deleted from Jira) the cache will be updated accordingly.

Although not recommended, the cache can be disabled/refreshed by checking the newly added option “Refresh cached data at next display“ from the gadget settings.

The changes are live. Please let me know if you see any improvements.

We will continue by improving the Kanban Velocity gadget.

Danut.

[Danut M [StonikByte]]

Hi @{5b2d16639ba6d02662e34fbd}

We introduced a data caching for Cycle Time Trend and Kanban Velocity Gadgets.

Please let me know if you see any improvements.

Danut.

[Danut M [StonikByte]]

This is fixed for both Cloud (1.1.24-AC) and Server (1.12.0) by introducing data caching.