Closed robert-bryson closed 2 years ago
We've increased the ingress-nginx timeout, which we believe was the source of the problem. However, we have yet to verify this as we need to complete load-testing first.
Most recent result from the above query... Not sure how to interpret it yet.
I checked this link, https://catalog-stage-datagov.app.cloud.gov/dataset/?page=1800
It still takes a long time to load, but it actually loads.
I checked this link, https://catalog-stage-datagov.app.cloud.gov/dataset/?page=1800
It still takes a long time to load, but it actually loads.
On staging it timed 20+ seconds for most of deep pagination pages (> 15000). On prod it was about 40 seconds, probably due to prod Solr was busy at re-indexing.
We rarely see the 500 error. We will mark resolved.
Solr returns a 504 Gateway Time-out with query https://default-solr-2e001a6400465b1e-solrcloud.54a8e908-f9ea-49ea-8891-d927fd60261e.ssb.data.gov/solr/ckan/select?q=%3A&start=311111 (or any query that hits deep in the pagination).
How to reproduce
Expected behavior
Responds successfully and quickly
Actual behavior
504 Gateway Time-out error
Sketch
This was surfaced with deep paginations of the cloud.gov catalog site (like this for example) taking a long time to return (around 1min).
Probably related to https://github.com/GSA/datagov-deploy/issues/3642