| chandankumar | sean-k-mooney, amoralej Hello, Please have a look at https://review.opendev.org/c/openstack/watcher-dashboard/+/958209: Add skip action feature with microversion support, I have added the integration test for skip workflow here https://review.opendev.org/c/openstack/watcher-dashboard/+/976594 and here is the job result: https://zuul.opendev.org/t/openstack/build/caf3bc578b1947e3a25e67101d8969ab and generated | 08:40 |
|---|---|---|
| chandankumar | screenshot from the test: https://4e653b44258ae0ef8e8c-431c4675dbf321a8fbd861f82b52f6e9.ssl.cf2.rackcdn.com/openstack/caf3bc578b1947e3a25e67101d8969ab/controller/logs/playwright/screenshots/watcher_dashboard.test.integration.test_playwright_skip_action_workflow.SkipActionWorkflowTests.test_skip_action_workflow/index.html | 08:40 |
| chandankumar | I want to get this one merged in this release itself. thank you! | 08:40 |
| opendevreview | Joan Gilabert proposed openstack/watcher master: Prepare to use openstacksdk instead of novaclient https://review.opendev.org/c/openstack/watcher/+/974710 | 08:48 |
| opendevreview | Joan Gilabert proposed openstack/watcher master: Complete migration from novaclient to openstacksdk https://review.opendev.org/c/openstack/watcher/+/974924 | 08:48 |
| opendevreview | Joan Gilabert proposed openstack/watcher master: Remove usage of novaclient from Watcher https://review.opendev.org/c/openstack/watcher/+/974925 | 08:48 |
| opendevreview | Joan Gilabert proposed openstack/watcher master: Remove nova_helper retries for openstacksdk params https://review.opendev.org/c/openstack/watcher/+/975498 | 08:48 |
| amoralej | sorry chandankumar, I will test that | 08:56 |
| opendevreview | Joan Gilabert proposed openstack/watcher master: Prepare to use openstacksdk instead of novaclient https://review.opendev.org/c/openstack/watcher/+/974710 | 11:43 |
| opendevreview | Joan Gilabert proposed openstack/watcher master: Complete migration from novaclient to openstacksdk https://review.opendev.org/c/openstack/watcher/+/974924 | 11:43 |
| opendevreview | Joan Gilabert proposed openstack/watcher master: Remove usage of novaclient from Watcher https://review.opendev.org/c/openstack/watcher/+/974925 | 11:43 |
| opendevreview | Joan Gilabert proposed openstack/watcher master: Remove nova_helper retries for openstacksdk params https://review.opendev.org/c/openstack/watcher/+/975498 | 11:43 |
| opendevreview | Joan Gilabert proposed openstack/watcher master: Prepare to use openstacksdk instead of novaclient https://review.opendev.org/c/openstack/watcher/+/974710 | 14:11 |
| opendevreview | Joan Gilabert proposed openstack/watcher master: Complete migration from novaclient to openstacksdk https://review.opendev.org/c/openstack/watcher/+/974924 | 14:11 |
| opendevreview | Joan Gilabert proposed openstack/watcher master: Remove usage of novaclient from Watcher https://review.opendev.org/c/openstack/watcher/+/974925 | 14:11 |
| opendevreview | Joan Gilabert proposed openstack/watcher master: Remove nova_helper retries for openstacksdk params https://review.opendev.org/c/openstack/watcher/+/975498 | 14:11 |
| opendevreview | Joan Gilabert proposed openstack/watcher master: Prepare to use openstacksdk instead of novaclient https://review.opendev.org/c/openstack/watcher/+/974710 | 16:35 |
| opendevreview | Joan Gilabert proposed openstack/watcher master: Complete migration from novaclient to openstacksdk https://review.opendev.org/c/openstack/watcher/+/974924 | 16:36 |
| opendevreview | Joan Gilabert proposed openstack/watcher master: Remove usage of novaclient from Watcher https://review.opendev.org/c/openstack/watcher/+/974925 | 16:36 |
| opendevreview | Joan Gilabert proposed openstack/watcher master: Remove nova_helper retries for openstacksdk params https://review.opendev.org/c/openstack/watcher/+/975498 | 16:36 |
| amoralej | Hi, as part of the instrumentalization of watcher, i'm trying to create a way to measure the size of the DataModel. I've being testing https://paste.centos.org/view/8dfe6dbd (with the help of claude) to get the size by using sys.getsizeof transversing the networkx graph and objects inside. I'm comparing that with the reported size by https://pympler.readthedocs.io/en/stable/library/asizeof.html | 16:55 |
| amoralej | and I'm getting very different values | 16:55 |
| amoralej | for a small model get_memory_size reports 1557 bytes, while asizeof reports 7424 | 16:57 |
| amoralej | you have any idea about what may be a good way of getting the size of complex objects ? | 16:58 |
| opendevreview | sean mooney proposed openstack/watcher master: Add retry_on_deadlock decorators to missing database methods https://review.opendev.org/c/openstack/watcher/+/976293 | 17:09 |
| opendevreview | sean mooney proposed openstack/watcher master: Add regression test for retry_on_deadlock decorator https://review.opendev.org/c/openstack/watcher/+/977248 | 17:09 |
| sean-k-mooney | amoralej: i woudl preer not to add any depencies for doing that if we can avoid it | 17:10 |
| amoralej | yes, that's why i didn't want to use pympler or any lib like that and testing the method in https://paste.centos.org/view/8dfe6dbd | 17:11 |
| sean-k-mooney | amoralej: its not really somethign i have looked into | 17:11 |
| sean-k-mooney | maybe via the inspect module? | 17:12 |
| amoralej | and was using pympler to compare the result | 17:12 |
| sean-k-mooney | i kind of wantto remvoe the xml serialistion eventualy | 17:12 |
| sean-k-mooney | but perhaps we could jsut count the bytes of that as a proxy | 17:12 |
| amoralej | yes, i was also thinking on that, but my guess is that the size of the serialized to string content can be very different .. | 17:13 |
| amoralej | i can check and compare | 17:13 |
| amoralej | size_with_get_memory_size: 1557 | 17:16 |
| amoralej | pympler.asizeof: 7424 | 17:16 |
| amoralej | text_size: 396 | 17:16 |
| amoralej | all those are bytes | 17:17 |
| sean-k-mooney | yes it can but its a proxy | 17:31 |
| sean-k-mooney | i think if you use inspect you might eb able to just get teh size of the stack fram and or local variables | 17:31 |
| sean-k-mooney | you could try https://docs.python.org/3/library/tracemalloc.html | 17:34 |
| amoralej | yes, I've also been playing with that | 17:34 |
| amoralej | I'll keep working on it | 17:36 |
| opendevreview | Joan Gilabert proposed openstack/watcher master: Prepare to use openstacksdk instead of novaclient https://review.opendev.org/c/openstack/watcher/+/974710 | 17:37 |
| opendevreview | Joan Gilabert proposed openstack/watcher master: Complete migration from novaclient to openstacksdk https://review.opendev.org/c/openstack/watcher/+/974924 | 17:37 |
| opendevreview | Joan Gilabert proposed openstack/watcher master: Remove usage of novaclient from Watcher https://review.opendev.org/c/openstack/watcher/+/974925 | 17:37 |
| opendevreview | Joan Gilabert proposed openstack/watcher master: Remove nova_helper retries for openstacksdk params https://review.opendev.org/c/openstack/watcher/+/975498 | 17:37 |
| sean-k-mooney | amoralej: you could just look at the overall process usage before and after building or copying the model | 17:39 |
| sean-k-mooney | https://docs.python.org/3/library/resource.html#resource.getrusage | 17:39 |
| sean-k-mooney | that is almost freee as it just reading form proc i think | 17:40 |
| sean-k-mooney | as in for /proc/self/... | 17:40 |
| sean-k-mooney | that ownt tell you wnat is using it however | 17:41 |
| amoralej | yes, i checked with process.memory_info().rss | 17:43 |
| sean-k-mooney | amoralej: claude says this shoudl work https://paste.opendev.org/show/b13If3Y2f7DPX8tP4VW1/ | 17:43 |
| amoralej | but that's even more misleading in many cases | 17:43 |
| sean-k-mooney | it identifed an issue with hwo you were recusing | 17:43 |
| amoralej | lemme try that | 17:44 |
| sean-k-mooney | https://paste.opendev.org/show/bFQ4wQjaJTwH4fh4rKdX/ uses a stack to make it iteritive isntead ot avoid the recurssion limit | 17:45 |
| sean-k-mooney | those two examples will not count interened string that are storead at teh process level and shared ro c memory allcotin but of pure python object it shoudl count the size correctly | 17:48 |
| sean-k-mooney | i think that is more then good enough however assumign it works | 17:49 |
| amoralej | size_with_get_memory_size: 1557 | 17:53 |
| amoralej | pympler.asizeof: 7424 | 17:53 |
| amoralej | text_size: 396 | 17:53 |
| amoralej | model.deep_sizeof_report: 4272 | 17:53 |
| sean-k-mooney | are you consucting the model the same way in ech case | 17:54 |
| amoralej | yes, it's exactly the same model actually, just executing different methods on it | 17:54 |
| amoralej | need to leave now, i'll keep working on it, thanks for your help! | 17:54 |
| sean-k-mooney | i ugess as long as we choose oen way of doign hte messurement | 17:54 |
| sean-k-mooney | adn we are consitent that likely good enough | 17:55 |
| sean-k-mooney | chat to you tomorrow o/ | 17:55 |
| amoralej | yeah, i agree | 17:56 |
| opendevreview | Douglas Viroel proposed openstack/watcher master: Enable Applier parallel engine in native thread mode https://review.opendev.org/c/openstack/watcher/+/975522 | 18:08 |
| opendevreview | Merged openstack/watcher master: Use wrapper classes for novaclient objects https://review.opendev.org/c/openstack/watcher/+/972913 | 20:41 |
Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!