...
[16:05:00 CDT(-0500)] <EricDalquist> there is a background process that runs every 200ms which goes through that queue and calls interrupt on each worker until the worker eventually stops
[16:05:50 CDT(-0500)] <athena> EricDalquist: i think i've found the issue with the layout.json document. that URL wasn't behind the URL canonicalizing filter until this commit: https://github.com/Jasig/uPortal/commit/53d54bc66ea808414614d62a2a52175a3b167da0
[16:06:07 CDT(-0500)] <athena> er, wrong link - https://github.com/Jasig/uPortal/commit/6e7396b547a32ea2ecdf301e0f03de6db85c675d
[16:06:20 CDT(-0500)] <EricDalquist> ah
[16:06:29 CDT(-0500)] <EricDalquist> right that url shouldn't get canonicalized
[16:06:30 CDT(-0500)] <drewwills> i see... so if we had another process – or perhaps the same process – count the ones that won't die by fname, it could trigger the no-new-threads valve
[16:06:38 CDT(-0500)] <EricDalquist> yup
[16:06:47 CDT(-0500)] <EricDalquist> so drewwills there is no right answer here