Open drewmiranda-gl opened 5 days ago
When using the metrics exporter for graylog, the behavior of the following metrics has changed between Graylog 5.2 and 6.0:
gl_jvm_memory_total_max
gl_jvm_memory_non_heap_max
In Graylog 5.2 these displayed the configured max value for JVM heap. In Graylog 6.0 they show -1 or unlimited:
-1
Its not clear to me which of these is the correct value to return though.
server.conf
prometheus_exporter_enabled = true prometheus_exporter_bind_address = 0.0.0.0:9833 prometheus_exporter_mapping_file_path_custom = /usr/share/graylog-server/prometheus-exporter-mapping-custom.yml
Relevant metrics declared in prometheus-exporter-mapping-custom.yml
prometheus-exporter-mapping-custom.yml
metric_mappings: - metric_name: "jvm_memory_total_max" match_pattern: "jvm.memory.total.max" - metric_name: "jvm_memory_non_heap_max" match_pattern: "jvm.memory.non-heap.max"
6.0 includes updates to our metrics libraries. E.g. https://github.com/Graylog2/graylog2-server/pull/17142. That could be related.
When using the metrics exporter for graylog, the behavior of the following metrics has changed between Graylog 5.2 and 6.0:
gl_jvm_memory_total_max
gl_jvm_memory_non_heap_max
In Graylog 5.2 these displayed the configured max value for JVM heap. In Graylog 6.0 they show
-1
or unlimited:Its not clear to me which of these is the correct value to return though.
Expected Behavior
Current Behavior
Possible Solution
Steps to Reproduce (for bugs)
server.conf
Relevant metrics declared in
prometheus-exporter-mapping-custom.yml
Context
Your Environment