Regular went down checks
Opened this issue · 9 comments
I have several down events on check for only one minute.
Obsioulsy, the concerned website wasn't down at this moment.
In most case, I have socket hang up error and few Parse Error
It's similar of #216.
Any idea ?
I use the master version and have actually 136 checks configured.
Seeing a lot of ECONNRESET and socket hang ups (few times a day), making the tests flap, resulting in actually issues being missed.
Is this an uptime problem or a server problem (given the number of requests the server is handling a day it could well be, but want to be sure before I go back to the owner).
Now I've seen this issue and the related issue I'm not so sure.
We also don't really know if is an uptime or server issue.
Still investigating...
Same issue here. Any investigation going on?
Discover this log in dmesg:
[...]
[1014705.347074] type=1401 audit(1408437344.848:465): SELinux: unrecognized netlink message type=256 for sclass=30
[1015085.727817] type=1401 audit(1408437724.937:466): SELinux: unrecognized netlink message type=256 for sclass=30
[1015235.932512] type=1401 audit(1408437875.031:467): SELinux: unrecognized netlink message type=256 for sclass=30
[1015656.414921] type=1401 audit(1408438295.189:468): SELinux: unrecognized netlink message type=256 for sclass=30
[1015676.405604] type=1401 audit(1408438315.165:469): SELinux: unrecognized netlink message type=256 for sclass=30
Any relation with this bug ? I got some difficulties to find it out on Google search...
Found a similar issue on another projet: Automattic/knox#116
It seems to be solved by changing http.globalAgent.maxSockets settings. See here: Automattic/knox#116 (comment)
Still investigating.
@soullivaneuh What results from your Investigation in the last 8 days?
Absolutly nothing interesting...
I'm suffering the same problem, with 51 checks every 5 minutes
The SELinux: unrecognized netlink message type=256 for sclass=30 message is related to this glibc bug: https://sourceware.org/bugzilla/show_bug.cgi?id=15946