Details
-
Type:
Change Request
-
Status: Closed
-
Priority:
Trivial
-
Resolution: Fixed
-
Affects Version/s: 4.0.0rc2
-
Fix Version/s: 3.0.23rc1, 3.4.15rc1, 4.0.1rc1, 4.2.0alpha1, 4.2 (plan)
-
Component/s: Proxy (P), Server (S)
-
Labels:
-
Team:Team A
-
Sprint:Sprint 44
-
Story Points:0.25
Description
There is useful function zbx_mem_dump_stats which dumps various memory statistics, for example:
9941:20181001:110548.601 === memory statistics for configuration cache ===
9941:20181001:110548.602 free chunks of size >= 256 bytes: 1
9941:20181001:110548.603 min chunk size: 8114960 bytes
9941:20181001:110548.604 max chunk size: 8114960 bytes
9941:20181001:110548.605 memory of total size 8388232 bytes fragmented into 1796 chunks
9941:20181001:110548.605 of those, 8114960 bytes are in 1 free chunks
9941:20181001:110548.606 of those, 244552 bytes are in 1795 used chunks
9941:20181001:110548.606 ================================
However when out of memory error occurs it is not called.
Only this is available:
1397:20180920:030749.965 __mem_malloc: skipped 586 asked 248031720 skip_min 256 skip_max 222534240 1397:20180920:030749.966 file:vmware.c,line:87 zbx_mem_realloc(): out of memory (requested 248031720 bytes) 1397:20180920:030749.966 file:vmware.c,line:87 zbx_mem_realloc(): please increase VMwareCacheSize configuration parameter
Adding more statistics could help diagnosing out of memory errors in case there is fragmentation or something unusual.