Hello everyone,
We are currently using the following plugin to retrieve all AWS CloudWatch alarms:
/usr/lib/centreon/plugins/centreon_aws_cloudwatch_api.pl --plugin=cloud::aws::cloudwatch::plugin --mode=get-alarms --custommode='awscli' --aws-secret-key='***' --aws-access-key='***' --aws-role-arn='' --proxyurl='' --region='sa-east-1' --filter-alarm-name='' --warning-status='%{state_value} =~ /INSUFFICIENT_DATA/i' --critical-status='%{state_value} =~ /ALARM/i' --verbose
We have an active alarm, which can be confirmed when using the --debug option. The output indicates that the instance memory has <StateValue>ALARM</StateValue>. However, the final result is:
ok: status : skipped (no value(s))
My question is: could this be a potential bug in the plugin?
When running the native AWS CLI command:
aws cloudwatch describe-alarms --region sa-east-1 --state-value ALARM
all active alarms are correctly returned, while the plugin does not display them.
Below is the --debug output showing the alarm:
<member>
<TreatMissingData>breaching</TreatMissingData>
<AlarmConfigurationUpdatedTimestamp>2025-09-01T18:58:33.193Z</AlarmConfigurationUpdatedTimestamp>
<StateValue>ALARM</StateValue>
<Threshold>40.0</Threshold>
<StateReason>Threshold Crossed: no datapoints were received for 1 period and 1 missing datapoint was treated as tBreaching].</StateReason>
<InsufficientDataActions/>
<StateTransitionedTimestamp>2025-09-02T04:39:15.668Z</StateTransitionedTimestamp>
<AlarmActions/>
<StateUpdatedTimestamp>2025-09-02T04:39:15.668Z</StateUpdatedTimestamp>
<ComparisonOperator>GreaterThanThreshold</ComparisonOperator>
<AlarmName>Instance memory</AlarmName>
<EvaluationPeriods>1</EvaluationPeriods>
<StateReasonData>{"version":"1.0","queryDate":"2025-09-02T04:39:15.664+0000","period":300,"recentDatapoints":i],"threshold":40.0,"evaluatedDatapoints":a{"timestamp":"2025-09-02T04:34:00.000+0000"}]}</StateReasonData>
<ActionsEnabled>true</ActionsEnabled>
<DatapointsToAlarm>1</DatapointsToAlarm>
<Metrics>
<member>
<Expression>MAX(e1)</Expression>
<ReturnData>true</ReturnData>
<Label>Instance memory usage</Label>
<Id>e2</Id>
</member>
<member>
<Period>300</Period>
<Expression>SELECT AVG(mem_used_percent) FROM SCHEMA(CWAgent,InstanceId) GROUP BY InstanceId ORDER BY AVG() DESC</Expression>
<ReturnData>false</ReturnData>
<Label>Instance</Label>
<Id>e1</Id>
</member>
</Metrics>
<CreationId>87b60bf6-d984-4612-8e58-c77ed25e9809/1751060096124</CreationId>
<OKActions/>
<AlarmArn>arn:aws:cloudwatch:sa-east-1:074071149174:alarm:Instance memory</AlarmArn>
<Dimensions/>
</member>
</MetricAlarms>
</DescribeAlarmsResult>
<ResponseMetadata>
<RequestId>af6003e5-476b-400c-a06f-77ce382f0c99</RequestId>
</ResponseMetadata>
</DescribeAlarmsResponse>
2025-09-02 09:44:31,070 - MainThread - botocore.hooks - DEBUG - Event needs-retry.cloudwatch.DescribeAlarms: calling handler <botocore.retryhandler.RetryHandler object at 0x7f8741f3e650>
2025-09-02 09:44:31,071 - MainThread - botocore.retryhandler - DEBUG - No retry needed.
OK: 0 problem(s) detected | 'alerts'=0;;;0;
Command line: 'aws cloudwatch describe-alarms --region sa-east-1 --output json --debug'
ok: status : skipped (no value(s))
ok: status : skipped (no value(s))