...

• Events that are written by ONAP components.
• Propagation of transaction and invocation information between components.
• MDCs, Markers and other information that should be attached to log messages.
  • MDC = Mapped Diagnostic Context
• Human and machine-readable output format(s).
• Files, locations and other conventions. 

Java and Python are supported, but conventions may be implemented by other technologies like GO.   

Original AT&T ONAP Logging guidelines (pre amsterdam release) - for historical reference only: https://wiki.onap.org/download/attachments/1015849/ONAP%20application%20logging%20guidelines.pdf?api=v2

The Acumos logging specification follows this document at https://wiki.acumos.org/display/OAM/Log+Standards 

Logback reference: Logging Developer Guide#Logback.xml based on https://gerrit.onap.org/r/#/c/62405

Introduction

The purpose of ONAP logging is to capture information needed to operate, troubleshoot and report on the performance of the ONAP platform and its constituent components. Log records may be viewed and consumed directly by users and systems, indexed and loaded into a datastore, and used to compute metrics and generate reports. 

...

This document proposes conventions you can follow to generate conformant, indexable logging output from your component.

How to Log

ONAP prescribes conventions. The use of certain APIs and providers is recommended, but they are not mandatory. Most components log via EELF or SLF4J to a provider like Logback or Log4j.

Logging Specification Compliance

Supported Languages

Java (including Scala and Clojure - because they compile to the same bytecode) and Python are supported, but conventions may be implemented by other technologies like GO. 

Java

The library and aop wrapper are written in Java and will work for Clojure (thanks to a discussion with Shwetank Dave that reminded me of languages that compile to bytecode) and Scala as well. see: https://git.onap.org/logging-analytics/tree/reference

Clojure

See the java libraries

Scala

see the java libraries

Python

https://lists.onap.org/g/onap-discuss/topic/logging_python_logging/25307286?p=,,,20,0,0,0::recentpostdate%2Fsticky,,,20,2,0,25307286

How to Log

ONAP prescribes conventions. The use of certain APIs and providers is recommended, but they are not mandatory. Most components log via EELF or SLF4J to a provider like Logback or Log4j.

Logging Specification Compliance

For the casablanca release the logging specification has been updated and finalized as of June 2018.For the casablanca release there the logging specification has been updated and finalized as of June 2018.  Implementation of this specification is required but the method of implementation is optional based on each team's level of possible engagement.

...

Use a spring based Aspect library that emits Markers automatically around function calls and retrofit your code to log via Luke's SLF4J for internal log messages.

LoggingWithAOP48534506

Logging Library Location and Use

...

• Localization. 
• Error codes. 
• Generated wiki documentation. 

• Separate audit, metrics, security error and debug logs. 

EELF is a facade, so logging output is configured in two ways:

1. By selection of a logging provider such as Logback or Log4j, typically via the classpath. 
2. By way of a provider configuration document, typically logback.xml or log4j.xml. See Providers48534506.

SLF4J

SLF4J is a logging facade, and a humble masterpiece. It combines what's common to all major, modern Java logging providers into a single interface. This decouples the caller from the provider, and encourages the use of what's universal, familiar and proven. 

...

Logback is the most commonly used provider. It is generally configured by an XML document named logback.xml. See Configuration 48534506.

See HELM template https://git.onap.org/logging-analytics/tree/reference/provider/helm/logback

...

Log4j 2.X is somewhat less common than Logback, but equivalent. It is generally configured by an XML document named log4j.xml. See Configuration 48534506.

Log4j 1.X

Strongly discouraged from Beijing onwards, since 1.X is EOL, and since it does not support escaping, so its output may not be machine-readable. See https://logging.apache.org/log4j/1.2/.

...

An important aspect of this is analytics, which requires tracing of requests between components. In a large, distributed and scalable system such as ONAP this is critical to understanding behavior and performance. 

...

General 

It isn't the aim of this document to reiterate the basicsBest Practices, so advice here is general: 

• Use a logger. Consider using logging facade such as SLF4J or EELF. 
• Write log messages in English.
• Write meaningful messages. Consider what will be useful to consumers of logger output. 
• Use errorcodes to characterise exceptions.
• Log at the appropriate level. Be aware of the volume of logs that will be produced.
• Safeguard the information in exceptions, and ensure it is never lost.
• Use error codes to characterize exceptions. 
• Log in a machine-readable format. See Conventions.
• Log for analytics as well as troubleshooting.

...

• http://www.masterzen.fr/2013/01/13/the-10-commandments-of-logging/
• https://www.loggly.com/blog/how-to-write-effective-logs-for-remote-logging/
• And so on.

Standard Attributes

These are attributes common to all log messages.

...

They are either:

• Explicitly required by logging APIs:
  
  • Logger
  • Level
  • Message
  • Exception (note that exception is the only standard attribute that may routinely be empty).
• Implicitly derived by the logging provider:
  • Timestamp
  • Thread.

Which means you normally can't help but report them. 

See https://www.slf4j.org/api/org/slf4j/Logger.html and https://logback.qos.ch/manual/layouts.html#ClassicPatternLayout for their origins and use.

Logger Name

This indicates the name of the logger that logged the message.

In Java it is convention to name the logger after the class or package using that logger.

• In Java, report the class or package name.
• In Python, the class or source filename.

Most other languages will fit one of those patterns. 

Level

Severity, typically drawn from the enumeration {TRACE, DEBUG, INFO, WARN, ERROR}.

Think carefully about the information you report at each log level. The default log level is INFO.

Some loggers define non-standard levels, like FINE, FINER, WARNING, SEVERE, FATAL or CRITICAL. Use these judiciously, or avoid them.

Message

The free text payload of a log event. 

This is the most important item of information in most log messages. See 48534506 guidelines.

Internationalization

Diagnostic log messages generally do not need to be internationalized.

Parameterization

Parameterized messages allow serialization to be deferred until AFTER level threshold checks. This means the cost is never incurred for messages that won't be written. 

• Favor parameterized messages, especially for INFO and DEBUG logging.
• Perform expensive serialization in the #toString method of wrapper classes.

For example:

logger.debug("Relax - this won't hurt: {}", new ToStringWrapper(costlyToSerialize));

Exception

The error stacktrace, where applicable.

Log unabridged stacktraces upon error.

When rethrowing, ensure that frame information is not lost:

• By logging the original exception at the point where it was caught. 
• By setting the original exception as the cause when rethrowing.

Timestamp

Logged as an ISO8601 UTC datetime. Millisecond (or greater) precision is preferable.

For example:

2018-07-05T20:21:34.794Z

Offset timestamps are OK provided the offset is included. (In the above example, the "Z" is a shorthand indicating an offset of zero – UTC).

Thread

The name of the thread from which the log message was emitted.

Thread names don't necessarily convey useful information, and their reliability depends on the thread model implemented by different runtimes, but they are sometimes used in heuristic analysis.

Efficiency

There is tension between utility and efficiency. IO bandwidth is finite, and the cost of serialization can be significant, especially at higher diagnostic levels.

Methods and Line Numbers

Many loggers can use reflection to emit the originating (Java) method, and even individual line numbers.

This information is certainly useful, but very expensive. Most logging implementations recommend that this not be enabled in production.

Level Thresholds

Level indicates severity.

Logger output is typically filtered by logger and level. The default logging level is INFO, so particular consideration should be given to the efficiency of INFO-level logging.

When DEBUG-level logging is configured, it's probably for good reason, and a greater overhead is expected. Be aware that it's not unusual for DEBUG logging to be left enabled inadvertently, however.

WARN and ERROR-level messages are of higher value, and comparatively rare, so their cost is less of a concern.

Conditionals

A common pattern is to place conditionals around (expensive) serialization.

For example:

if (logger.isDebugEnabled()) {
    logger.debug("But this WILL hurt: " + costlyToSerialize);
}

Parameterized logging is preferable.

Context

MDCs

A Mapped Diagnostic Context (MDC) allows an arbitrary string-valued attribute to be attached to a Java thread via a ThreadLocal variable. The MDC's value is then emitted with each message logged by that thread. The set of MDCs associated with a log message is serialized as unordered name-value pairs (see 48534506).

A good discussion of MDCs can be found at https://logback.qos.ch/manual/mdc.html. 

Example

From Luke Parker's call graph work in https://git.onap.org/logging-analytics/tree/reference/logging-slf4j-demo

LogEntry(markers=ENTRY, logger=ComponentAlpha, requestID=eb3e0dc2-6c3c-4bb7-8ed6-e5cc4ec7aad2, invocationID=06c815ef-5969-45cc-b319-d0dbcde89329, timestamp=Tue May 08 04:23:27 AEST 2018)

Mapped Diagnostic Context Table

• Must be set as early in invocation as possible. 
• Must be unset on exit. 
• keep in sync with https://wiki.acumos.org/display/OAM/Log+Standards

Context

MDCs

A Mapped Diagnostic Context (MDC) allows an arbitrary string-valued attribute to be attached to a Java thread via a ThreadLocal variable. The MDC's value is then emitted with each message logged by that thread. The set of MDCs associated with a log message is serialized as unordered name-value pairs (see Text Output).

A good discussion of MDCs can be found at https://logback.qos.ch/manual/mdc.html. 

Example

From Luke Parker's call graph work in https://git.onap.org/logging-analytics/tree/reference/logging-slf4j-demo

LogEntry(markers=ENTRY, logger=ComponentAlpha, requestID=eb3e0dc2-6c3c-4bb7-8ed6-e5cc4ec7aad2, invocationID=06c815ef-5969-45cc-b319-d0dbcde89329, timestamp=Tue May 08 04:23:27 AEST 2018)

...

• Must be set as early in invocation as possible. 
• Must be unset on exit. 

Logging

Via SLF4J:

import java.util.UUID;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import org.slf4j.MDC;
// ...
final Logger logger = LoggerFactory.getLogger(this.getClass());
MDC.put("SomeUUID", UUID.randomUUID().toString());
try {
    logger.info("This message will have a UUID-valued 'SomeUUID' MDC attached.");
    // ...
}
finally {
    MDC.clear();
}

...

• It's only a few calls. 
• It can be largely abstracted in the case of EELF logging.
  
  

MDC - InstanceID

(formerly InstanceUUID)

If known, this field contains a universally unique identifier used to differentiate between multiple instances of the same (named) log writing service/application. Its value is set at instance creation time (and read by it, e.g., at start/initialization time from the environment). This value should be picked up by the component instance from its configuration file and subsequently used to enable differentiation of log records created by multiple, locally load balanced ONAP component or subcomponent instances that are otherwise identically configured.

...

This field indicates the high level status of the request. It must have the value COMPLETE when the request is successful and ERROR when there is a failure.  And INPROGRESS for states between the two.

Discussion: status/response/severity relationship

...

Used as valuable URI - to annnote invoke marker

Review in terms of Marker 48534506 - INVOKE - possiblly add INVOKE-return - to filter reporting

...

TBD: cover off discussion on reducing log files to two (DEBUG/rest) for C* release

MDC -

...

TargetElement

VNF/PNF context dependent - on CRUD operations of VNF/PNFs

...

• There is considerable duplication:
  • BeginTimestamp, EndTimestamp, ElapsedTime. These are all captured elsewhere (and ElapsedTime is even redundant within that triplet).
  • Server, ServerIPAddress, ServerFQDN, VirtualServiceName. Overkill. Should be one, plus optionally ClientIPAddress (or some variant thereof).
  • TargetEntity, TargetServiceName, not obviously different to similar attributes.
• There is junk:
  • Severity? Nagios codes?
  • ProcessKey?
  • All the stuff that's already grayed out in the table above.
  • People may defend these individually, maybe vigorously, but they're domain-specific:
    • That absolutely doesn't mean they can't be used.
    • Beats configuration allows ad hoc contexts to be indexed.
    • But perhaps they don't belong in this kind of spec.
• Redundant attributes *do* matter, because:
  • Populating and propagating everything prescribed by the guide approaches being prohibitive. People won't do it, and people *don't* do it. 
  • If something might be in one of several attributes then that's worse than it being in just one.
• That means:
• We're left with only two MANDATORY attributes, necessary to build invocations graphs:
  • RequestID - top-level transactions. 
  • InvocationID - inter-component invocations.
• And a minimal number of OPTIONAL descriptive attributes: ServiceInstanceID, InstanceUUIDInstanceID, Server, StatusCode, ResponseCode, ResponseDescription.
• Those are the ones we need to document clearly, support in APIs, etc.
• That's <=10, a manageable number. 
• And again, that matters because if the number isn't manageable, people won't (and don't) comply. 

...

2018-07-05T20:21:34.794Z	http-nio-8080-exec-2	INFO	org.onap.demo.logging.ApplicationService	InstanceUUIDInstanceID=ede7dd52-91e8-45ce-9406-fbafd17a7d4c, RequestID=f9d8bb0f-4b4b-4700-9853-d3b79d861c5b, ServiceName=/logging-demo/rest/health/health, InvocationID=8f4c1f1d-5b32-4981-b658-e5992f28e6c8, InvokeTimestamp=2018-07-05T20:21:26.617Z, PartnerName=, ClientIPAddress=0:0:0:0:0:0:0:1, ServerFQDN=localhost			ENTRY	
2018-07-05T20:22:09.268Z	http-nio-8080-exec-2	INFO	org.onap.demo.logging.ApplicationService	ResponseCode=, InstanceUUIDInstanceID=ede7dd52-91e8-45ce-9406-fbafd17a7d4c, RequestID=f9d8bb0f-4b4b-4700-9853-d3b79d861c5b, ServiceName=/logging-demo/rest/health/health, ResponseDescription=, InvocationID=8f4c1f1d-5b32-4981-b658-e5992f28e6c8, Severity=, InvokeTimestamp=2018-07-05T20:21:26.617Z, PartnerName=, ClientIPAddress=0:0:0:0:0:0:0:1, ServerFQDN=localhost, StatusCode=			EXIT	

...

This should accompany INVOKE when the invocation is synchronous.

...

SLF4J:

public static final Marker INVOKE_SYNCHRONOUS;
static {
    INVOKE_SYNCHRONOUS = MarkerFactory.getMarker("INVOKE");
    INVOKE_SYNCHRONOUS.add(MarkerFactory.getMarker("SYNCHRONOUS"));
}
// ...

// Generate and report invocation ID. 

final String invocationID = UUID.randomUUID().toString();
MDC.put(MDC_INVOCATION_ID, invocationID);
try {
    logger.debug(INVOKE_SYNCHRONOUS, "Invoking synchronously ... ");
}
finally {
    MDC.remove(MDC_INVOCATION_ID);
}

// Pass invocationID as HTTP X-InvocationID header.

callDownstreamSystem(invocationID, ... );

EELF example of SYNCHRONOUS reporting, without changing published APIs. 

...

Error Codes

Errorcodes Error codes are reported as MDCs. 

TODO: add to table

Exceptions should be accompanied by an errrorcodeerror code. Typically this is achieved by incorporating errorcodes error codes into your exception hierarchy and error handling. ONAP components generally do not share this kind of code, though EELF defines a marker interface (meaning it has no methods) EELFResolvableErrorEnum.

A common convention is for errorcodes error codes to have two components:

1. A prefix, which identifies the origin of the error. 
2. A suffix, which identifies the kind of error.

...

1. Logs should be human-readable (within reason). 
2. Shipper and indexing performance and durability depends on logs that can be parsed quickly and reliably.
3. Consistency means fewer shipping and indexing rules are required.

Text Output

TODO: 20190115 - do not take the example in this section until I reverify it in terms of the reworked spec example in 

https://gerrit.onap.org/r/#/c/62405/20/reference/logging-kubernetes/logdemonode/charts/logdemonode/resources/config/logback.

...

xml

...

ONAP needs to strike a balance between human-readable and machine-readable logs. This means:

...

An error log is required for EELF-enabled components, and is intended to capture info, warn, error and fatal conditions sensed (“exception handled”) by the software components.

This includes previous logs that went to application.log

Debug Log

A debug log is optional for EELF-enabled components, and is intended to capture whatever data may be needed to debug and correct abnormal conditions of the application.

...

Console logging may also be present, and is intended to capture “system/infrastructure” records. That is stdout and stderr assigned to a single “engine.out” file in a directory configurable (e.g. as an environment/shell variable) by operations personnel.

Application Log (deprecated)

see example in https://git.onap.org/oom/tree/kubernetes/portal/charts/portal-sdk/resources/config/deliveries/properties/ONAPPORTALSDK/logback.xml

We no longer support this 5th log file - see error.log 

Log File Locations

There are several locations where logs are available on the host, on the nfs share and in each application and filebeat container

...

Logs on each filebeat docker container sidecar - /var/log/onap

...

...

...

...

What's New

(Including what WILL be new in v1.2  / R2). 

...