As I'm no longer using LogDNA I can't spare the time to maintain this library sorry. You are welcome to run your own fork, however there is another similar library that is being maintained. https://github.com/zileo-net/logback-logdna I haven't used it though.
If somebody else wants to adopt this library, please don't hesiate to ask.
End of notice.
LogDNA is a hosted logging platform: https://logdna.com
This small library provides an appender for logback (a popular logging subsystem for the JVM).
The appender pushes log entries to LogDNA via HTTPS
Logback's thread-bound storage, the MDC, is sent to LogDNA as metadata to be indexed and then searchable. (See screenshot, and more below)
Logback uses XML file in known locations with the most common being classpath:/logback.xml
This is what a logback.xml might look like:
<?xml version="1.0" encoding="UTF-8"?>
<configuration>
<!-- Take both of these LOGDNA appenders -->
<!-- this one is a plain old HTTP transport -->
<appender name="LOGDNA-HTTP" class="net._95point2.utils.LogDNAAppender">
<appName>LogDNA-Logback-Test</appName>
<ingestKey>${LOGDNA_INGEST_KEY}</ingestKey>
<includeStacktrace>true</includeStacktrace>
</appender>
<!-- ... and this one should be attached to the root -->
<appender name="LOGDNA" class="ch.qos.logback.classic.AsyncAppender">
<appender-ref ref="LOGDNA-HTTP" />
</appender>
<appender name="CONSOLE" class="ch.qos.logback.core.ConsoleAppender">
<encoder>
<pattern>%m%n</pattern>
<charset>utf8</charset>
</encoder>
</appender>
<root level="INFO">
<appender-ref ref="CONSOLE" />
<appender-ref ref="LOGDNA" /><!-- the async one -->
</root>
</configuration>
This setup uses an asynchronous wrapper to deal with log shipping on another thread and has some sensible options for dealing with buffer management. Read more here
- Add a link to this project's repository in your pom
project > repositories >
<repository>
<snapshots>
<enabled>false</enabled>
</snapshots>
<id>bintray-robshep-oss</id>
<name>bintray</name>
<url>https://dl.bintray.com/robshep/oss</url>
</repository>
- Add the dependency
project > dependencies >
<dependency>
<groupId>net._95point2.utils</groupId>
<artifactId>logback-logdna</artifactId>
<version>1.1.0</version>
</dependency>
Or just download the Jar and it's dependencies from https://github.com/robshep/logback-logdna/releases
You can go to town on most other logback configurations but the LogDNA only has a couple of settings (shown as the default)
<appName>LogDNA-Logback-Test</appName>
set this for good log management in LogDNA<ingestKey>${LOGDNA_INGEST_KEY}</ingestKey>
signup to LogDNA and find this in your account profile<includeStacktrace>true</includeStacktrace>
this library can send multiline stacktraces (see image) - Raw syslog transport cannot<sendMDC>true</sendMDC>
copies over logback's Mapped Diagnostic Context (MDC) as LogDNA Metadata which are then indexed and searchable.
- The log line displays the thread, the logger (class) and the message
- LogDNA's metadata is populated with the logger as an indexable/searchable property.
- The HTTP Transport is done by the very lightweight DavidWebb REST Library and so doesn't introduce bulky dependencies
The combination of logback's MDC and LogDNA's metadata support is pretty powerful and means you can correlate web-requests, or a userId, or something else that happens in a thread in your application.
For example, doing this anywhere in your application...
MDC.put("customerId", "C-1");
MDC.put("requestId", "cafebabe1");
... means that you can then go and search for say, that customer in logDNA like this:
meta.customerId:"C-1"