Hi, after heavy utilization over time password log...
# talk-kratos
f
Hi, after heavy utilization over time password login flows for Kratos 0.5, query below started to take even 25 seconds to yield results. What could be the issue? Is flow Id table purged after usage or there is s limitation in 0.5 version. Could you shed some light on this please? Query in question:
SELECT selfservice_login_flow_methods.config, selfservice_login_flow_methods.created_at, selfservice_login_flow_methods.id, selfservice_login_flow_methods.method, selfservice_login_flow_methods.selfservice_login_flow_id, selfservice_login_flow_methods.updated_at FROM selfservice_login_flow_methods AS selfservice_login_flow_methods WHERE selfservice_login_flow_id = $1
h
yes this should become better in ory kratos 0.7!
and 0.8
f
If update isn't an instant fix, should i purge the table with flows? Is there any temp fix regarding this? Thanks for answering 🖖
h
yeah you can purge those tables
f
Out of curiosity, can it suddenly increase in processing time from 1/ms up to 4-20/seconds in a span of 10 minutes? Is there this critical point where that could be overflowing? Postgress 9.6
h
hm I think the tables were missing some indices in kratos 0.5
which can cause these spikes
f
Great, much appreciated 🙏
h
any time 🙂