Load average on AWS does not divide by number of cores

[rafaeldff]

All "processors" on the output of cat /proc/cpuinfo have the same "core id". It seems each of them is a part of a hyperthreaded core.

Currently, riemann-health behaves as if the machine has a single processor, always dividing the load average by 1. I don't know if it's possible to get the actual number of cores assigned to the VM, but there is probably a better estimate than 1.

E.g, here is the output of a 4 vCPU instance, reporting all "processors" as having core id = 2:

processor    : 0
vendor_id    : GenuineIntel
cpu family    : 6
model        : 45
model name    : Intel(R) Xeon(R) CPU E5-2670 0 @ 2.60GHz
stepping    : 7
microcode    : 0x70b
cpu MHz        : 2599.998
cache size    : 20480 KB
physical id    : 0
siblings    : 4
core id        : 2
cpu cores    : 1
apicid        : 5
initial apicid    : 5
fpu        : yes
fpu_exception    : yes
cpuid level    : 13
wp        : yes
flags        : fpu de tsc msr pae cx8 sep cmov pat clflush mmx fxsr sse sse2 ss ht syscall nx lm constant_tsc rep_good nopl nonstop_tsc pni pclmulqdq ssse3 cx16 pcid sse4_1 sse4_2 x2apic popcnt tsc_deadline_timer aes hypervisor lahf_lm ida arat epb xsaveopt pln pts dtherm
bogomips    : 5199.99
clflush size    : 64
cache_alignment    : 64
address sizes    : 46 bits physical, 48 bits virtual
power management:

processor    : 1
vendor_id    : GenuineIntel
cpu family    : 6
model        : 45
model name    : Intel(R) Xeon(R) CPU E5-2670 0 @ 2.60GHz
stepping    : 7
microcode    : 0x70b
cpu MHz        : 2599.998
cache size    : 20480 KB
physical id    : 0
siblings    : 4
core id        : 2
cpu cores    : 1
apicid        : 5
initial apicid    : 5
fpu        : yes
fpu_exception    : yes
cpuid level    : 13
wp        : yes
flags        : fpu de tsc msr pae cx8 sep cmov pat clflush mmx fxsr sse sse2 ss ht syscall nx lm constant_tsc rep_good nopl nonstop_tsc pni pclmulqdq ssse3 cx16 pcid sse4_1 sse4_2 x2apic popcnt tsc_deadline_timer aes hypervisor lahf_lm ida arat epb xsaveopt pln pts dtherm
bogomips    : 5199.99
clflush size    : 64
cache_alignment    : 64
address sizes    : 46 bits physical, 48 bits virtual
power management:

processor    : 2
vendor_id    : GenuineIntel
cpu family    : 6
model        : 45
model name    : Intel(R) Xeon(R) CPU E5-2670 0 @ 2.60GHz
stepping    : 7
microcode    : 0x70b
cpu MHz        : 2599.998
cache size    : 20480 KB
physical id    : 0
siblings    : 4
core id        : 2
cpu cores    : 1
apicid        : 5
initial apicid    : 5
fpu        : yes
fpu_exception    : yes
cpuid level    : 13
wp        : yes
flags        : fpu de tsc msr pae cx8 sep cmov pat clflush mmx fxsr sse sse2 ss ht syscall nx lm constant_tsc rep_good nopl nonstop_tsc pni pclmulqdq ssse3 cx16 pcid sse4_1 sse4_2 x2apic popcnt tsc_deadline_timer aes hypervisor lahf_lm ida arat epb xsaveopt pln pts dtherm
bogomips    : 5199.99
clflush size    : 64
cache_alignment    : 64
address sizes    : 46 bits physical, 48 bits virtual
power management:

processor    : 3
vendor_id    : GenuineIntel
cpu family    : 6
model        : 45
model name    : Intel(R) Xeon(R) CPU E5-2670 0 @ 2.60GHz
stepping    : 7
microcode    : 0x70b
cpu MHz        : 2599.998
cache size    : 20480 KB
physical id    : 0
siblings    : 4
core id        : 2
cpu cores    : 1
apicid        : 5
initial apicid    : 5
fpu        : yes
fpu_exception    : yes
cpuid level    : 13
wp        : yes
flags        : fpu de tsc msr pae cx8 sep cmov pat clflush mmx fxsr sse sse2 ss ht syscall nx lm constant_tsc rep_good nopl nonstop_tsc pni pclmulqdq ssse3 cx16 pcid sse4_1 sse4_2 x2apic popcnt tsc_deadline_timer aes hypervisor lahf_lm ida arat epb xsaveopt pln pts dtherm
bogomips    : 5199.99
clflush size    : 64
cache_alignment    : 64
address sizes    : 46 bits physical, 48 bits virtual
power management:

[jamtur01]

@rafaeldff - The branch I've just pushed should address this issue I think. Is it possible for you to test - here's a gem build with the patch applied:

https://www.dropbox.com/sh/bm8g5e8s94eu66n/AACTNDFX8Vb7SwUNyU4jqmKIa?dl=0

[adamdyga]

I've faced this issue too. Any chance the fix gets to regular release? It's difficult to set thresholds for CPU load if the load is not reported on per-core basis.

[jamtur01]

@adamdyga Does the fix address the issue? I'm still waiting for feedback.

[adamdyga]

@jamtur01 to me it works like a charm ;)