MessageLogging policy

What

One of the best ways to track down problems in the API runtime environment is to log messages. You can attach and configure a MessageLogging policy on your API to log custom messages to syslog.

Samples

Syslog

<MessageLogging name="LogToSyslog">
  <Syslog>
    <Message>[3f509b58 tag="{organization.name}.{apiproxy.name}.{environment.name}"] Weather request for WOEID {request.queryparam.w}.</Message>
    <Host>logs-01.loggly.com</Host>
    <Port>514</Port>
    <Protocol>TCP</Protocol>
    <FormatMessage>true</FormatMessage>
  </Syslog>
  <logLevel>ALERT</logLevel>
</MessageLogging>

A common usage of the MessageLogging policy type is to log to a syslog account. When configured for syslog, an API proxy will forward log messages from Apigee to a remote syslog server. You must already have a syslog server available. If not, public log management services, such a Splunk, Sumo Logic, and Loggly, are available. see Configuring third-party log management services.

For example, imagine that you need to log information about each request message that your API receives from consumer apps. The value 3f509b58 represents a key value specific to the loggly service. If you have a loggly account, substitute your loggly key. The log message that is generated will be populated with four values: the organization, API proxy, and environment name associated with the transaction, along with the value for a query parameter on the request message.

Note: Host and Port elements cannot refer to variables. They must contain static values.

Syslog over TLS/SSL

<MessageLogging name="LogToSyslog">
  <Syslog>
    <Message>[3f509b58 tag="{organization.name}.{apiproxy.name}.{environment.name}"] Weather request for WOEID {request.queryparam.w}.</Message>
    <Host>logs-01.loggly.com</Host>
    <Port>6514</Port>
    <Protocol>TCP</Protocol>
    <FormatMessage>true</FormatMessage>
    <SSLInfo>
        <Enabled>true</Enabled>
    </SSLInfo>
  </Syslog>
  <logLevel>WARN</logLevel>
</MessageLogging>

You can send messages to third-party message logging providers over TLS/SSL by adding the <SSLInfo> block.

File rotation: size

<MessageLogging name="LogPolicy">
  <File>
    <Message>This is a test message. Message id : {request.header.messageid}</Message>
      <FileName>test.log</FileName>
      <FileRotationOptions rotateFileOnStartup="true">
        <FileRotationType>SIZE</FileRotationType>
        <MaxFileSizeInMB>10</MaxFileSizeInMB>
        <MaxFilesToRetain>10</MaxFilesToRetain>
      </FileRotationOptions>
  </File>
  <logLevel>ERROR</logLevel>
</MessageLogging>

File rotation based on file size.

File rotation: time

<MessageLogging name="LogPolicy">
  <File>
    <Message>This is a test message. Message id : {request.header.messageid}</Message>
    <FileName>test.log</FileName>
    <FileRotationOptions rotateFileOnStartup="true">
      <FileRotationType>TIME</FileRotationType>
      <RotationFrequency unit="minute">10</RotationFrequency>
      <MaxFilesToRetain>10</MaxFilesToRetain>
    </FileRotationOptions>
  </File>
  <logLevel>ERROR</logLevel>
</MessageLogging>

File rotation based on time.

File rotation: time & size

<MessageLogging name="LogPolicy">
  <File>
    <Message>This is a test message. Message id : {request.header.messageid}</Message>
    <FileName>test.log</FileName>
    <FileRotationOptions rotateFileOnStartup="true">
      <FileRotationType>TIME_SIZE</FileRotationType>
      <MaxFileSizeInMB>10</MaxFileSizeInMB>
      <MaxFilesToRetain>10</MaxFilesToRetain>
      <RotationFrequency unit="minute">10</RotationFrequency>
    </FileRotationOptions>
  </File>
  <logLevel>ERROR</logLevel>
</MessageLogging>

File rotation based on time and size.

Stream-enabled

<MessageLogging name="LogPolicy">
  <File>
  ....
  ....
  </File>
  <BufferMessage>true</BufferMessage>
</MessageLogging>

Stream-enabled message logging

Note: If HTTP streaming is enabled in an API proxy, neither request nor response messages will be buffered by processing pipeline. When you need to log parsed message content, set BufferMessage to true.

---

Element reference

Use the elements in the table below to configure the MessageLogging policy type.

To send syslog to Splunk, Sumo Logic, or Loggly, see Configuring third-party log management services.

Note: The name attribute for this policy is restricted to these characters: A-Z0-9._\-$ %. However, the Apigee UI enforces additional restrictions, such as automatically removing characters that are not alphanumeric.

Field Name	Field Description
Message	Build the message to be sent to the syslog, combining text with variables to capture the information you want. See the Samples. The <Message> element supports the dynamic string substitution feature called message templating. Note: Response variables will not be available in PostClientFlow following an Error Flow. Use message variables to log response information for both error and success situations. See also Usage notes.
Host	The hostname or IP address of the server where the syslog should be sent. If you don't include this element, the default is localhost.
Port	Port where the syslog is running. If you don't include this element, the default is 514.
Protocol	TCP or UDP (default). While UDP is more performant, the TCP protocol guarantees message log delivery to the syslog server. For sending syslog messages over TLS/SSL, only TCP is supported.
FormatMessage	true or false (default) Optional, but <FormatMessage>true</FormatMessage> is required for use with Loggly. This element lets you control the format of Apigee-generated content prepended to the message. If set to true, the syslog message is prepended by a fixed number of characters, which lets you filter out that information from messages. Here's an example for the fixed format: <14>1 2016-02-23T09:24:39.039+0000 e49cd3a9-4cf6-48a7-abb9-7ftfe4d97d00 Apigee - - - Message starts here The Apigee-generated information includes: <14> - A priority score (see the Syslog Protocol) based on the log level and facility level of the message. 1 - The current syslog version. Date with a UTC offset (UTC = +0000). Message processor UUID. "Apigee - - - " If set to false (default), the message is not prepended with those fixed characters. Note: If you set <FormatMessage>false</FormatMessage>, the log message still includes the priority score and the date. To format the log message to only include the message body, set <PayloadOnly>true</PayloadOnly>. See PayloadOnly.
PayloadOnly	true or false (default) This element sets the format of Apigee-generated messages to contain only the body of the syslog message, without the prepended characters specified by FormatMessage. If you don't include this element or leave it empty, the default value is false. See FormatMessage. Note: setting <PayloadOnly>true</PayloadOnly> overrides the setting of FormatMessage.
SSLInfo	Lets you log messages through SSL/TLS. Use with sub-element <Enabled>true</Enabled>. If you don't include this element or leave it empty, the default value is false (no TLS/SSL). <SSLInfo> <Enabled>true</Enabled> </SSLInfo> You can configure the <SSLInfo> tag the same as you can on a TargetEndpoint, including enabling two-way TLS/SSL, as described in API proxy configuration reference. Only the TCP protocol is supported.
logLevel	Optional. Valid values: INFO (default), ALERT, WARN, ERROR Set a specific level of information to be included in the message log. If you're using the FormatMessage element (setting it to true), your logLevel setting affects the calculated priority score (the number inside the angle brackets) in the Apigee-generated information prepended to the message.

Schemas

Sample: See our GitHub repository samples for the most recent schemas.

---

Usage notes

When attaching a MessageLogging policy to an API proxy flow, consider placing it in the ProxyEndpoint response, in a special flow called PostClientFlow. The PostClientFlow executes after the response is sent to the requesting client, which ensures that all metrics are available for logging. For details on using PostClientFlow, see API proxy configuration reference.

Note: It's important to note that the response flow variable is not available in the PostClientFlow once processing enters the error state. Instead, you can use the message variable to set values in the MessageLogging policy in the PostClientFlow, ensuring that they will be set whether or not the error state was the previous context. For more information and examples, see "Uses for message variables".

The PostClientFlow is special in two ways:

1. It only executed as part of the response flow.
2. It is the only flow executed after the proxy enters the error state.

Because it is executed regardless of whether the proxy succeeded or failed, you can put MessageLogging policies in the PostClientFlow and be guaranteed that they always execute.

The following Trace image shows a MessageLogging policy executing as part of the PostClientFlow, after the DefaultFaultRule executes:

In this example, the Verify API Key policy caused the fault because of an invalid key.

Shown below is the ProxyEndpoint definition that includes the PostClientFlow:

<ProxyEndpoint name="default">
  ...
  <PostClientFlow>
    <Response>
      <Step>
        <Name>Message-Logging-1</Name>
      </Step>
    </Response>
  </PostClientFlow>
  ...
</ProxyEndpoint>

Apigee logs messages as simple text, and you can configure logging to include variables, such as the date and time when the request or response was received, the user identity on the request, the source IP address from which the request was sent, and so on. Apigee logs message asynchronously, meaning that no latency that might be caused by blocking callouts is introduced to your API.

The MessageLogging policy writes logged messages in memory to a buffer. The message logger reads messages from the buffer and then writes to the destination that you configure. Each destination has its own buffer.

Note: Proxy calls succeed when logging fails

Because message logging is first written to buffer, the API proxy will continue successful execution even if message logging ultimately fails (for example, if there's a connection failure to the external syslog provider).

If the write rate to the buffer increases beyond the read rate, the buffer overflows and logging will fail. If this happens, you might find a message containing the following in the log file:

Log message size exceeded. Increase the max message size setting

Increase the max.log.message.size.in.kb property (default value = 128 KB) in the message-logging.properties file.

Note: The response message variables in Apigee are not available from the Error Flow. These variables are also not available in PostClientFlow if the preceding flow was the Error Flow. If you want to log response information from the PostClientFlow, use the message object. You can use this object to get at headers and other information from the response whether or not there was an error. See Message variables for more information and an example.

Default values for variables in message template

Default values can be specified for each variable in message template separately. For example, if the variable request.header.id cannot be resolved, then its value is replaced with the value unknown.

<Message>This is a test message. id = {request.header.id:unknown}</Message>

A common default value can be specified for all the unresolved variables by setting the defaultVariableValue attribute on the Message element:

<Message defaultVariableValue="unknown">This is a test message. id = {request.header.id}</Message>

Note: The default variable value cannot contain spaces.

Configuring third-party log management services

The MessageLogging policy lets you send syslog messages to third-party log management services, such as Splunk, Sumo Logic, and Loggly. If you want to send syslog to one of those services, see that service's documentation to configure the service's host, port, and protocol, then set the Syslog element on this policy accordingly.

See the following documentation for third-party log management configuration:

• Splunk (select the product version)
  Also see this Apigee Community post: https://community.apigee.com/content/kbentry/13298/log-messages-into-splunk.html
• Sumo Logic
  • Also see this Apigee Community post: https://community.apigee.com/questions/5226/setting-up-logging-with-sumo-logic-which-host-shou.html
  • For a complete example using Sumo Logic as the logging service, see the following Apigee Community post. The solution uses a single JavaScript policy to make HTTP POST requests to Sumo Logic HTTP Source Collector: https://community.apigee.com/articles/32286/logging-to-sumo-logic-using-javascript-and-http.html
• Loggly
  When using Loggly, <FormatMessage>true</FormatMessage> is required in the policy as a child of the <Syslog> element.
  Also see this Apigee Community post for more information about message logging to Loggly: https://community.apigee.com/content/kbentry/14798/log-messages-into-loggly.html

Error reference

This section describes the fault codes and error messages that are returned and fault variables that are set by Apigee when this policy triggers an error. This information is important to know if you are developing fault rules to handle faults. To learn more, see What you need to know about policy errors and Handling faults.

Runtime errors

These errors can occur when the policy executes.

Fault code	HTTP status	Cause
steps.messagelogging.StepDefinitionExecutionFailed	500	See fault string.

Deployment errors

These errors can occur when you deploy a proxy containing this policy.

Need help resolving an error? Click in the Fix column for detailed troubleshooting information.

Error name	Cause	Fix
InvalidProtocol	The deployment of the MessageLogging policy can fail with this error if the protocol specified within the <Protocol> element is not valid. The valid protocols are TCP and UDP. For sending syslog messages over TLS/SSL, only TCP is supported.
InvalidPort	The deployment of the MessageLogging policy can fail with this error if the port number is not specified within the <Port> element or if it is not valid. The port number must be an integer greater than zero.

Fault variables

These variables are set when a runtime error occurs. For more information, see What you need to know about policy errors.

Note: Proxy calls succeed when logging fails

There's a difference between policy errors and message logging errors. The flow variables here are populated only when the policy itself fails, not when message logging fails. Because message logging is first written to buffer, the API proxy will continue successful execution even if message logging ultimately fails (for example, if there's a connection failure to the external syslog provider). Be sure to check your logs on a regular basis to make sure logging is happening as expected.

Variables	Where	Example
fault.name="fault_name"	fault_name is the name of the fault, as listed in the Runtime errors table above. The fault name is the last part of the fault code.	fault.name Matches "StepDefinitionExecutionFailed"
messagelogging.policy_name.failed	policy_name is the user-specified name of the policy that threw the fault.	messagelogging.ML-LogMessages.failed = true

Example error response

Note: For error handling, the best practice is to trap the errorcode part of the error response. Do not rely on the text in the faultstring, because it could change.

{  
   "fault":{
      "detail":{
         "errorcode":"steps.messagelogging.StepDefinitionExecutionFailed"
      },
      "faultstring":"Execution failed"
   }
}

Example fault rule

<FaultRule name="MessageLogging">
    <Step>
        <Name>ML-LogMessages</Name>
        <Condition>(fault.name Matches "StepDefinitionExecutionFailed") </Condition>
    </Step>
    <Condition>(messagelogging.ML-LogMessages.failed = true) </Condition>
</FaultRule>

---

Flow variables

The following variables are populated on policy failure.

• messagelogging.failed
• messagelogging.{stepdefinition-name}.failed

Note: Proxy calls succeed when logging fails

There's a difference between policy errors and message logging errors. The flow variables here are populated only when the policy itself fails, not when message logging fails. Because message logging is first written to buffer, the API proxy will continue successful execution even if message logging ultimately fails (for example, if there's a connection failure to the external syslog provider). Be sure to check your logs on a regular basis to make sure logging is happening as expected.

---

Related topics

• Variables exposed by Apigee: Variables reference