Using bulk metrics for the log counters reduces largely the likelihood
of blocking the Heka pipeline. Instead of injecting (x services
* y levels) metric messages, the filter injects only one big message.
This changes also updates the configuration of the metric_collector
service to deserialize the bulk metric to support alarms on log
counters.
Change-Id: Icb71fd6faa4191795c0470ecc24aeafd25794f42
Closes-Bug: #1643280
This patch makes hostname an optional field. Currently here are metrics
that have no hostname:
- Some metrics provided by hypervisor_stats:
- total_free_disk_GB
- total_free_ram_MB
- total_free_vcpus
- total_used_disk_GB
- total_used_ram_MB
- total_used_vcpus
- total_running_instances
- total_running_tasks
- all metrics collected by check_openstack_api
- all metrics collected by http_check
Change-Id: I4b1078ddf6ef510ae2c95ae6937b28f007d88bea
This patch makes hostname an optional field. Currently here are metrics
that have no hostname:
- Some metrics provided by hypervisor_stats:
- total_free_disk_GB
- total_free_ram_MB
- total_free_vcpus
- total_used_disk_GB
- total_used_ram_MB
- total_used_vcpus
- total_running_instances
- total_running_tasks
- all metrics collected by check_openstack_api
- all metrics collected by http_check
Change-Id: Ic503b48e995170efd2b87c9385750fe920e2e25a
The logs_counter filter needs to discard the log messages that are not in the
current interval otherwise the aggregated metrics can be incorrect.
Change-Id: I40e08b9a0bba6d9b8c5042a0fd16a6818a42b9c9
Closes-Bug: #1549873
This change adds a new filter that processes OpenStack logs and adds a new
field that contains the number of logs per service and level to produce a
new metric. The metric is the number of logs written per second.
Change-Id: Ie55ee55be66550e9f97fdfd9b7df75af2d90fd19
Implements: blueprint detect-spike-of-errors-in-the-logs