Zipkin, from Twitter

A distributed tracing system



Scala 2.9 (JDK6 or JDK7) is required.

Install with git clone && cd zipkin or download Zipkin to your preferred directory and unzip.

Run (in separate bash windows):


Now you can access the Zipkin UI at http://localhost:8080/

Advanced Setup

By default, Zipkin runs on SQLite because it doesn't require additional setup. However, Zipkin is typically used with other databases (most often Cassandra) in production. For a quickstart with Cassandra, see the Ubuntu Quickstart and Mac Quickstart guides. These will help you get a somewhat more advanced Zipkin installation running on a single machine so that you can experiment with it.

For larger deployments, Zipkin works with several other services, and you may want to apply additional configuration. Some of those services and configurations are described below.

Feeding trace data to Zipkin

The next step after installation is to collect trace data to view in Zipkin. To do this, you must interface with the collector daemon to record trace data. There are several libraries to make this easier to do in different environments. Twitter uses Finagle; external libraries (currently for Python, REST, node, and Java) are listed in the wiki; and there is also a Ruby gem and Ruby Thrift client.

Running the Daemons with Other Databases

The default Zipkin collector configuration is zipkin-collector-service/config/collector-dev.scala and the default Zipkin query configuration is zipkin-query-service/config/query-dev.scala. These are set up to work with SQLite by default. To use Zipkin with a different SQL database, we recommend changing those configurations as described in the SQL guide.

For NoSQL databases, you have a few options. To run Zipkin with Cassandra for example, you can copy the Cassandra configuration into the default configuration, change the bin scripts to run the Cassandra configuration, or run the full daemon commands that the bin scripts are invoking:

bin/sbt 'project zipkin-collector-service' 'run -f zipkin-collector-service/config/collector-cassandra.scala'
bin/sbt 'project zipkin-query-service' 'run -f zipkin-query-service/config/query-cassandra.scala'

The bin/web script does not access the database so you do not need to change anything there.

Storage: Cassandra

Zipkin runs on SQLite by default to make it easier to test out of the box. Running Zipkin on SQLite requires no additional setup. However, in production, Zipkin is most commonly used with Cassandra for the collector's storage. There are also plugins for Redis, HBase, and other SQL databases.

  1. See Cassandra's site for instructions on how to start a cluster.
  2. Use the Zipkin Cassandra schema attached to this project. You can create the schema with the following command: cassandra-cli -host localhost -port 9160 -f zipkin-cassandra/src/schema/cassandra-schema.txt

Coordination: ZooKeeper

Zipkin can use ZooKeeper for coordination. That's where we store the server side sample rate and register the servers.

  1. See ZooKeeper's site for instructions on how to install it.

Collecting Data: Scribe

Scribe is the logging framework we use at Twitter to transport the trace data. There are several other ways to tell Zipkin what trace data to collect; in particular, if you are just trying out Zipkin you can skip this step entirely and point the ZipkinTracer directly at the collector.

To use Scribe with Zipkin, you need to set up a network store that points to the Zipkin collector daemon. A Scribe store for Zipkin might look something like this:


If you don't want to hardcode the IP address of your collector there are a few options. One is to use an internal DNS entry for the collectors so that you only have one place to change the addresses when you add or remove collectors. Alternatively, if you want to get all fancy you can use a modified version of Scribe that picks up the collectors via ZooKeeper. When each collector starts up it adds itself to ZooKeeper and when a collector shuts down it is automatically removed. The modified Scribe gets notified when the set of collectors change. To use this mode you change remote_host in the configuration to zk://zookeeper-hostname:2181/scribe/zipkin or something similar.

We're hoping that others might add non-Scribe transports for the tracing data; there is no reason why Scribe has to be the only one.

Zipkin UI

The UI is a standard Rails 3 app.

  1. Update config with your ZooKeeper server. This is used to find the query daemons.
  2. Deploy to a suitable Rails 3 app server. For testing you can simply do bundle install && bundle exec rails server.

zipkin-tracer gem

The zipkin-tracer gem adds tracing to a Rails application through the use of a Rack Handler. In

  use ZipkinTracer::RackHandler

If the application's static assets are served through Rails, those requests will be traced.