Metric Sink
Overview
The Unified Assurance Metric Sink microservice is part of the microservice metric pipeline. The Metric Sink is responsible for taking metrics that have been normalized and inserting them into the Metric database.
Prerequisites
-
A microservices cluster must be setup. Refer to Microservice Cluster Setup.
-
Apache Pulsar must be installed. Refer to Apache Pulsar microservice.
Setup
su - assure1
export NAMESPACE=a1-zone1-pri
export WEBFQDN=<Primary Presentation Web FQDN>
a1helm install metric-sink assure1/metric-sink -n $NAMESPACE --set global.imageRegistry=$WEBFQDN
Default Configuration
Name | Value | Possible Values | Notes |
---|---|---|---|
LOG_LEVEL | INFO | FATAL, ERROR, WARN, INFO, DEBUG | Logging level used by application. |
STREAM_INPUT | persistent://assure1/metric/sink | Text, 255 characters | Apache Pulsar topic path. Topic at end of path may be any text value. |
Configurations can be changed by passing the values to the a1helm install
prefixed with the configData parent key.
Example of setting the log level to DEBUG
a1helm install ... --set configData.LOG_LEVEL=DEBUG
Autoscaling
For microservice scaling options, please refer to the autoscaling docs.
On top of the standard autoscaling options, this microservice supports the following additional configurations:
Name | Value | Possible Values | Notes |
---|---|---|---|
thresholds.backlogSize | 1000 | Integer | The number of items that need to be in the backlog before the auto-scaling starts additional processes. |
thresholds.totalMetricsProcessed | 400 | Integer | Total metrics processed by the microservice. If the average of total metrics processed in five minutes exceeds the threshold, pods will be scaled. |
Example of changing the auto-scaling value
To change the thresholds used for this functionality, pass the following to the installation command:
a1helm install ... --set autoscaling.thresholds.backlogSize=2000
Schema
Metric-sink is expecting metric measurements in the following schema example:
{
"@timestamp": "2021-05-06T18:40:00.000Z",
"measurement": "iot_test",
"tags": {
"host": "some_host",
"instance": "some_instance"
},
"fields": {
"availability": 1,
"value": 123456789000,
"utilization": 0.1,
"other": 1
}
}
The operator is responsible for defining the proper format, although metric-sink will do its best to match the current Metric schema. For example, it will clean up slash marks (), replace dashes (-) and spaces with underscores (_) and capitalize letters as needed.
Debugging
The metric-sink consists of two components: metric-sink and telegraf. To see logs for each of the components, run the following command (replacing appropriate parameters).
For metric-sink logs:
a1k logs <Pod Name> -n <Namespace> metric-sink
For telegraf logs:
a1k logs <Pod Name> -n <Namespace> telegraf
Note:
The <Pod Name>
and <Namespace>
values are available in the get pods
command:
a1k get pods --all-namespaces
Microservice self-metrics
The Unified Assurance Metric Sink microservice exposes the following self-metrics to Prometheus.
Metric Name | Type | Description |
---|---|---|
metric_sink_backlog_size | Gauge | Number of items pending in the backlog |
total_metrics_processed | Counter | Number of metrics processed |
processing_time_per_metric | Gauge | processing time per metric from receiving it to sending it to acknowledging it |