You can set up CloudWatch to receive your application log and add an alarm to warn you of possible errors. This allows you to have more visible notifications whenever your deployed applications encounter exceptions. The following sections help you understand and learn about the configuration of CloudWatch logging and alarm setup.
Deployment of CloudWatch logging
By default, the AWS Blu Age Runtime contains a logging file named
logback-cloudwatch.yml. This file is referenced in the
application-main.yml file, but this reference is commented
out.
# logging:
# config: classpath:logback-cloudwatch.xmlBoth files are in the config folder, and by uncommenting the above lines, the feature can be activated. CloudWatch logging can be configured, as explained in the following sections.
Configuration of CloudWatch logging
The default logback-cloudwatch.xml file has the following
contents.
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE configuration>
<configuration>
<appender name="console" class="ch.qos.logback.core.ConsoleAppender">
<encoder>
<pattern>%date{yyyy-MM-dd HH:mm:ss.SSS,UTC} %level --- [%thread{15}] %logger{40} : %msg%n%xThrowable</pattern>
</encoder>
</appender>
<appender name="cloudwatch" class="com.netfective.bluage.runtime.cloudwatchlogger.CloudWatchAppender">
<logGroup>BluAgeRuntimeOnEC2-Logs</logGroup>
<logStream>%date{yyyy-MM-dd,UTC}.%instanceId.%uuid</logStream>
<layout>
<pattern>%date{yyyy-MM-dd HH:mm:ss.SSS,UTC} %level --- [%thread{15}] %logger{40} : %msg%n%xThrowable</pattern>
</layout>
<appender-ref ref="console" />
</appender>
<root level="INFO">
<appender-ref ref="cloudwatch" />
</root>
</configuration>Everything outside the <appender name="cloudwatch"/> element is standard logback configuration. There are two appenders in this file: a console appender to send logs to the console and a CloudWatch appender to send logs to CloudWatch.
The level attribute in the root element specifies the
logging level of the entire application.
The required values inside the tag <appender name="cloudwatch"/> are:
-
<logGroup/>:Sets the name of the log group in CloudWatch. If the value is not specified it defaults to
BluAgeRuntimeOnEC2-Logs. If the log group doesn’t exist it will be created automatically. This behavior can be changed through configuration, which is covered below. -
<logStream/>: Sets the name of the logStream (inside of the log group) in CloudWatch.
Optional values:
-
<region/>: Overrides the Region that the log stream will be written to. By default, logs go to the same Region as the EC2 instance.
-
<layout/>: The pattern the log messages will use.
-
<maxbatchsize/>: The maximum number of log messages to send to CloudWatch per operation.
-
<maxbatchtimemillis/>: The time in milliseconds to allow for CloudWatch logs to be written.
-
<maxqueuewaittimemillis/>: The time in milliseconds to try to insert requests in the internal log queue.
-
<internalqueuesize/>: The maximum size of the internal queue.
-
<createlogdests/>: Create log group and log stream if they don't exist.
-
<initialwaittimemillis/>: The amount of time that you want the thread to sleep on startup. This initial wait allows for an initial accrual of logs.
-
<maxeventmessagesize/>: The maximum size of a log event. Logs that exceed this size won’t be sent.
-
<truncateeventmessages/>: Truncate messages that are too long.
-
<printrejectedevents/>: Enable the emergency appender.
CloudWatch setup
In order for the above configuration to correctly push logs to CloudWatch, update your Amazon EC2 IAM instance profile role to grant it additional permissions for the `BluAgeRuntimeOnEC2-Logs` log group and its log streams:
-
logs:CreateLogStream -
logs:DescribeLogStreams -
logs:CreateLogGroup -
logs:PutLogEvents -
logs:DescribeLogGroups
Alarm setup
Thanks to CloudWatch logs, you can then configure different metrics and alarms, depending on your application and your needs. Specifically, you can set up proactive alarms for usage alerts, in order to be warned in the case of errors that might put your application in a grace period (and in the end, prevent it from working at all). To achieve this, you can add a metric concerning the "Error C5001" string in the logs, which highlights errors in the connection to the AWS Blu Age control system. You can then define an alarm that reacts to this metric.
