EKS demo: create-slo.sh fails, as no services available to cloudwatch #65
Description
This is using the EKS demo version of the repo.
The setup fails at at the create-slo.sh stage.
The reason it fails is because there are no services available for the list-services command to list:
++ aws application-signals list-services --endpoint https://application-signals.ap-southeast-1.api.aws --region ap-southeast-1 --cli-input-json '{
"StartTime": 1729037340,
"EndTime": 1729123740
}' --output json --query '(ServiceSummaries[?KeyAttributes.Name=='\''pet-clinic-frontend-java'\''].KeyAttributes)[0]'
+ SERVICE_KEY_ATTRIBUTES=null
+ echo null
nullHere is the full output of the script.
Admin:~/environment/application-signals-demo/scripts/eks/appsignals/one-step (main) $ ../create-slo.sh unicorn-pet-rentals ap-southeast-1
++ dirname ../create-slo.sh
+ cd ..
+ CLUSTER_NAME=unicorn-pet-rentals
+ REGION=ap-southeast-1
+ SERVICE_NAME=pet-clinic-frontend-java
+ ENDPOINT=https://application-signals.ap-southeast-1.api.aws
+ echo 'Creating Service Level Objectives'
Creating Service Level Objectives
++ date +%s
+ end_time=1729123740
+ end_time_as_int=1729123740
+ start_time_as_int=1729037340
+ LIST_SERVICES_REQUEST=slo/inputRequest/ListServices/listServices.json
++ sed -e 's|"StartTime": .*|"StartTime": 1729037340,|' -e 's|"EndTime": .*|"EndTime": 1729123740|' slo/inputRequest/ListServices/listServices.json
+ LIST_SERVICES_REQUEST_WITH_CORRECT_INPUT='{
"StartTime": 1729037340,
"EndTime": 1729123740
}'
++ aws application-signals list-services --endpoint https://application-signals.ap-southeast-1.api.aws --region ap-southeast-1 --cli-input-json '{
"StartTime": 1729037340,
"EndTime": 1729123740
}' --output json --query '(ServiceSummaries[?KeyAttributes.Name=='\''pet-clinic-frontend-java'\''].KeyAttributes)[0]'
+ SERVICE_KEY_ATTRIBUTES=null
+ echo null
null
+ '[' null = '' ']'
+ KEY_ATTRIBUTES=null
+ CREATE_SLO_REQUEST_1=slo/inputRequest/CreateServiceLevelObjective/getOwner99Availability.json
++ sed 's|"KeyAttributes": .*|"KeyAttributes": null,|' slo/inputRequest/CreateServiceLevelObjective/getOwner99Availability.json
+ CREATE_SLO_REQUEST_WITH_CORRECT_SERVICE_ARN_1='{
"Name": "Availability for Searching an Owner",
"Description": "Availability larger than 99 for Get Owner operation",
"SliConfig": {
"SliMetricConfig": {
"KeyAttributes": null,
"OperationName": "GET /api/customer/owners",
"MetricType": "AVAILABILITY",
"PeriodSeconds": 60
},
"MetricThreshold": 99.0,
"ComparisonOperator": "GreaterThan"
},
"Goal": {
"Interval": {
"RollingInterval": {
"Duration": 1,
"DurationUnit": "DAY"
}
},
"AttainmentGoal": 99.9,
"WarningThreshold": 60.0
}
}'
+ err=0
+ for i in {1..5}
++ aws application-signals create-service-level-objective --endpoint https://application-signals.ap-southeast-1.api.aws --cli-input-json '{
"Name": "Availability for Searching an Owner",
"Description": "Availability larger than 99 for Get Owner operation",
"SliConfig": {
"SliMetricConfig": {
"KeyAttributes": null,
"OperationName": "GET /api/customer/owners",
"MetricType": "AVAILABILITY",
"PeriodSeconds": 60
},
"MetricThreshold": 99.0,
"ComparisonOperator": "GreaterThan"
},
"Goal": {
"Interval": {
"RollingInterval": {
"Duration": 1,
"DurationUnit": "DAY"
}
},
"AttainmentGoal": 99.9,
"WarningThreshold": 60.0
}
}' --no-cli-pager --region ap-southeast-1
+ output='
Parameter validation failed:
Invalid type for parameter SliConfig.SliMetricConfig.KeyAttributes, value: None, type: <class '\''NoneType'\''>, valid types: <class '\''dict'\''>'
+ err=252
+ grep InvalidParameterValue
+ echo '
Parameter validation failed:
Invalid type for parameter SliConfig.SliMetricConfig.KeyAttributes, value: None, type: <class '\''NoneType'\''>, valid types: <class '\''dict'\''>'
+ break
+ check_if_loop_failed_and_exit 252 'There was an error creating an SLO - GetOwner99Availability, exiting'
+ '[' 252 -ne 0 ']'
+ echo There was an error creating an SLO - GetOwner99Availability, exiting
There was an error creating an SLO - GetOwner99Availability, exiting
+ exit 1For some reason, the CloudWatch SLO feature is not able to recognise the EKS cluster as a service.
EKS:
CloudWatch:
Someone else reported this issue back in April. Has this repo been tested recently?
Thanks