6 Policy Control Function Alerts
This section includes information about alerts for PCF.
Table 6-1 Common Alerts
Alert Name | Description | Severity |
---|---|---|
PCF_SERVICES_DOWN | Alert if any PCF service down for 5mins for given namespace in AlertRules file | Critical |
IngressErrorRateAbove10PercentPerPod | Alert if ingress error rate on each pod above 10% | Critical |
Table 6-2 SM Service Alerts
Alert Name | Description | Severity |
---|---|---|
SMTrafficRateAboveThreshold | Alert if Ingress traffic on SM service reaches 90% of max MPS in 2mins | Major |
SMIngressErrorRateAbove10Percent | Alert if Ingress transaction error rate exceeds 10% of all SM transactions in last 24 hours | Critical |
SMEgressErrorRateAbove1Percent | Alert if Egress transaction error rate exceeds 1% of all SM transactions in last 24 hours | Minor |
Table 6-3 Diameter Connector Alerts
Alert Name | Description | Severity |
---|---|---|
DiamTrafficRateAboveThreshold | Alert if Diameter Connector traffic reaches 90% of max MPS | Major |
DiamIngressErrorRateAbove10Percent | Alert if error rate exceeds 10% of all Diameter transactions in last 24 hours | Critical |
DiamEgressErrorRateAbove1Percent | Alert if Egress transaction error rate exceeds 1% of all Diameter transactions | Minor |
Table 6-4 User Service - UDR Alerts
Alert Name | Description | Severity |
---|---|---|
PcfUdrIngressTrafficRateAboveThreshold | Alert if Ingress traffic from UDR reaches 90% of max MPS | Major |
PcfUdrEgressErrorRateAbove10Percent | Alert if error rate exceeds 10% of all UDR transactions | Critical |
Table 6-5 User Service - CHF Alerts
Alert Name | Description | Severity |
---|---|---|
PcfChfIngressTrafficRateAboveThreshold | Alert if Ingress traffic from CHF reaches 90% of max MPS | Major |
PcfChfEgressErrorRateAbove10Percent | Alert if error rate exceeds 10% of all CHF transactions | Critical |
Table 6-6 PolicyDS Service Alerts
Alert Name | Description | Severity |
---|---|---|
PolicyDsIngressTrafficRateAboveThreshold | Alert if Ingress traffic reaches 90% of max MPS | Major |
PolicyDsIngressErrorRateAbove10Percent | Alert if Ingress error rate exceeds 10% of all PolicyDS transactions | Critical |
PolicyDsEgressErrorRateAbove1Percent | Alert if Egress error rate exceeds 10% of all PolicyDS transactions | Minor |
PCF Alert Configuration
This section describes the Measurement based Alert rules configuration for PCF. The Alert Manager uses the Prometheus measurements values as reported by microservices in conditions under alert rules to trigger alerts.
PCF Alert Configuration
Note:
- The alertmanager and prometheus tools should run in Oracle CNE namespace, for example, occne-infra.
- Alert file is packaged with PCF Custom Templates. The PCF Templates.zip file can be downloaded from OHC. Unzip the PCF Templates.zip file to get PcfAlertRules.yaml file.
- Edit the value of the following parameters in
thePcfAlertRules.yaml file before following the procedure for
configuring the alerts:
- [ 90% of Max MPS].
For Example, if the value of Max MPS is 10000, set [ 90% of Max MPS] as 9000 in yaml file as follows:
sum(rate(ocpm_ingress_request_total{servicename_3gpp="npcf-smpolicycontrol"}[2m])) >=9000
-
kubernetes_namespace.
For Example,
If PCF is deployed at more than one site, set kubernetes_namespace in yaml file as follows:expr: up{kubernetes_namespace=~"pcf|ocpcf"} == 0
If PCF is deployed at only one site, set kubernetes_namespace in yaml file as follows:
expr: up{kubernetes_namespace="pcf"}==0
- [ 90% of Max MPS].
To Configure PCF alerts in Prometheus:
- Find the config map to configure alerts in
prometheus server by executing the following command:
where, <Namespace> is the prometheus server namespace used in helm install command.kubectl get configmap -n <Namespace>
For Example, assuming prometheus server is under occne-infra namespace, execute the following command to find the config map:
0utput: occne-prometheus-server 4 46dkubectl get configmaps -n occne-infra | grep prometheus-server
- Take Backup of current config map
of prometheus server by executing the following command:
where, <Name> is the prometheus config map name used in helm install command.kubectl get configmaps <Name> -o yaml -n <Namespace> > /tmp/t_mapConfig.yaml
- Check if alertspcf is present in the t_mapConfig.yaml file by
executing the following command:
cat /tmp/t_mapConfig.yaml | grep alertspcf
- If alertspcf is present, delete the
alertspcf entry from the t_mapConfig.yaml file, by
executing the following command:
sed -i '/etc\/config\/alertspcf/d' /tmp/t_mapConfig.yaml
Note:
This command should be executed only once. - If alertspcf is not present, add the alertspcf entry in
the t_mapConfig.yaml file by executing the following command:
sed -i '/rule_files:/a\ \- /etc/config/alertspcf' /tmp/t_mapConfig.yaml
Note:
This command should be executed only once. - Reload the config map with the
modifed file by executing the following command:
kubectl replace configmap <Name> -f /tmp/t_mapConfig.yaml
- Add PcfAlertRules.yaml file
into prometheus config map by executing the following command :
where, <PATH> is the location of the PcfAlertRules.yaml file.kubectl patch configmap <Name> -n <Namespace> --type merge --patch "$(cat <PATH>/PcfAlertRules.yaml)"
- Restart prometheus-server pod.
- Verify the alerts in prometheus GUI. Below screenshot displays the PCF
alerts: