Kong all worker CPU 100%

[yuanst-627]

Our production environment is using 100% CPU for all worker processes.

Export the flamegraph as follows

./samples/lj-lua-stacks.sxx --skip-badvars -x 621952 > a.bt
The results are as follows
[root@api-prd-api02 stapxx]# cat a.bt
0x7f17d4f3916a
T:@/usr/local/share/lua/5.1/kong/router.lua:1485
@/usr/local/share/lua/5.1/kong/router.lua:1485
@/usr/local/share/lua/5.1/kong/runloop/handler.lua:1091
@/usr/local/share/lua/5.1/kong/init.lua:752
=access_by_lua(nginx-kong.conf:257):0
2118
0x7f17d4f39164
T:@/usr/local/share/lua/5.1/kong/router.lua:1485
@/usr/local/share/lua/5.1/kong/router.lua:1485
@/usr/local/share/lua/5.1/kong/runloop/handler.lua:1091
@/usr/local/share/lua/5.1/kong/init.lua:752
=access_by_lua(nginx-kong.conf:257):0
391

kong tag : 1.3.1

Can someone help me out here? thanks

[ProBrian]

Hi @yuanst-627 , I noticed that you mentioned kong tag: 1.3.1, do you mean version here? If not, could you provide more information: what version of kong are your running? kong ce or kong ee? are you running a heavy workload? are there any logs with warning or error s, or is your kong service is not working functional correctly?

[yuanst-627]

I've observed this issue. Is there a possible connection?
openresty/luajit2#42

[chronolaw]

I noticed this line "/usr/local/share/lua/5.1/kong/router.lua". Since kong gateway 3.x it will not exist.
I suggest to upgrade to latest version (3.7 or 3.4 LTS) and test again.