Help Troubleshooting significant delay in processing in SP

[Cantor, Scott]

Memcache, which is known to be unstable and unreliable (I mean the plugin) jumps out as the most obvious cause. That many transactions isn't going to cause anything unusual in the SP itself. A full on load test with the same NameID will blow it up badly, but not at that scale.

If the gap in the log is sitting inside a bunch of memcache calls, I don't see much reason to think it's anything else. Just stop using it. If you're sharing a socket, you have to be sharing a shibd instance, in which case memcache doesn't add anything anyway.

-- Scott