A field report on how running a Scala app on Google Container Engine, interfacing with several of Google Cloud's native services, has worked out for the engineers at the Norwegian startup Unacast.

We will discuss how we use Pub/Sub, BigQuery and Dataflow from Google Cloud's BigData feature set, and why we have chosen to use these services instead of setting up our own Kafka and Spark clusters.

Then we'll have a look at how we package, deploy and run our app as a pod in Kubernetes hosted on Google Container Engine using fancy ChatOps tricks. This part will also include a short overview of some of the logging and monitoring techniques we have applied, including Dropwizard Metrics + Datadog, Google Cloud Logging and Slack.

Last but not least I'll go into some detail of how it has been to use the Dataflow Java SDK from Scala, talking about how to define the Dataflow pipeline in somewhat functionally and what kind of customizations I had to do to make the experience more Scalaesque.

Torbjørn Vatn is currently a Senior Platform Engineer at the Norwegian startup Unacast where he programs Scala, Devs'n'Ops and brews a lot of coffee. He has 4-5 years experience as a full-time Scala developer working for both penger.no and BEKK Consulting.