Horizon pages are loaded too slowly when there are more than 500 elements per one page

[Paul Karikh]

On large datasets, when Horizon have to render more that 1000 elements in the table, pages are loaded too slowly. The more data - the more slow page will be loaded. So, after 10000 elements per page (10K users, for example).

Also, this issue is only with page without pagination. When page uses pagination, there is no problems with page load time.
If there are more than 500 elements in table - page load is about 30 seconds. 1000 elements in the table - page load time is about one minute. If more than 2000 elements - 2-3 minutes and more.

Pages:
1) /project/access_and_security/ (Security groups, key pairs, floating ips)
2) /project/network_topology/
3) /admin/routers/
4) /admin/networks/
5) /identity/users/
6) /admin/images/

[Timur Sufiev (tsufiev-x)]

Lowering to Medium because it's occurs only on specific configurations and does not completely prevent the User from doing his job.

[Timur Sufiev (tsufiev-x)]

Moving to 8.0 release due to lack of resources. SearchLight is going to be available in L-3, so we should take advantage of that.

[Timur Sufiev (tsufiev-x)]

Created a JIRA ticket for solving this kind of issues: https://mirantis.jira.com/browse/PROD-1985. Most changes are going to happen in Mitaka.

[Roman Rufanov (rrufanov)]

customer found on 6.0. Need fix.

[Vitaly Sedelnik (vsedelnik)]

Waiting for fix in 8.0

[Timur Sufiev (tsufiev-x)]

Just to set expectations here: this is not a bug, but rather a feature request (make table rendering in Horizon faster when there are a lot of objects) - there won't be single fix for this, more likely a bunch of different fixes for different tables + implementing pagination for the tables which don't have it yet. In some cases implementing pagination will require support on API side and in client. We won't be able to track the progress of this task in a single bug, please file a separate bugs for the tables we need to fix first.

[Denis Meltsaykin (dmeltsaykin)]

Setting this as Won't Fix for 6.0-updates, as this is not bug as per comment #6.

[Paul Karikh (pkarikh)]

Fix for /identity/users/ will use truncated flag from Keystone and api filtering.
Here is upstream Horizon patch for enabling `truncated flag` handling: https://review.openstack.org/#/c/250869/
Here is another upstream Horizon patch for adding api filtering for users. https://review.openstack.org/#/c/112744/15

[Timur Sufiev (tsufiev-x)]

There is one thing that may help us: https://blueprints.launchpad.net/horizon/+spec/cache-templates

[Timur Sufiev (tsufiev-x)]

Cached template loader patch was merged as part of bug 1524333.

[Timur Sufiev (tsufiev-x)]

Closing as Fix Committed, because in current state it cannot be used to track any new fixes. If other fixes for table rendering speedup are introduced, they must be tracked in separate bugs.

[Alexander Petrov (apetrov-n)]

We don't have the strict criteria for setting this issue as 'fix released'.
I tried it on 2000 users and page loading took about 1min 35sec.
But that is better than 3 min as mentioned in the bug's description.

Verified on MOS 8.0 Build 529

VERSION:
  feature_groups:
    - mirantis
  production: "docker"
  release: "8.0"
  api: "1.0"
  build_number: "529"
  build_id: "529"
  fuel-nailgun_sha: "baec8643ca624e52b37873f2dbd511c135d236d9"
  python-fuelclient_sha: "4f234669cfe88a9406f4e438b1e1f74f1ef484a5"
  fuel-agent_sha: "658be72c4b42d3e1436b86ac4567ab914bfb451b"
  fuel-nailgun-agent_sha: "b2bb466fd5bd92da614cdbd819d6999c510ebfb1"
  astute_sha: "b81577a5b7857c4be8748492bae1dec2fa89b446"
  fuel-library_sha: "e2d79330d5d708796330fac67722c21f85569b87"
  fuel-ostf_sha: "3bc76a63a9e7d195ff34eadc29552f4235fa6c52"
  fuel-mirror_sha: "fb45b80d7bee5899d931f926e5c9512e2b442749"
  fuelmenu_sha: "e071216cb214e34b4d861478033425ee6a54a3be"
  shotgun_sha: "63645dea384a37dde5c01d4f8905566978e5d906"
  network-checker_sha: "a43cf96cd9532f10794dce736350bf5bed350e9d"
  fuel-upgrade_sha: "616a7490ec7199f69759e97e42f9b97dfc87e85b"
  fuelmain_sha: "a365f05b903368225da3fea9aa42afc1d50dc9b4"