Thursday, September 1, 2011

TeamCity 7.0 EAP is Open: Agent Pools, Fail Build on Metric Change and More

So the codename for the next TeamCity major version (7.0) is to start with F. Not that much of a choice from Indian F-cities, so a tiny one, but most pleasantly sounding to the ears of our team: Faradi.

We are still to define 7.0 release date, but so far it seems like we will target it for the first quarter of 2012.

The good news are that you can try some of the features of upcoming 7.0 now: check the EAP release.
This build introduces features that will be further improved (based on your feedback!) to develop into solid features of 7.0.

As one of the focuses of 7.0 we plan various means to speed up your builds. This EAP starts the trend and introduces:
  • Agent Pools - allows to divide the agents into groups assigned to sets of projects, so you can dedicate more resources to specific projects easily;
  •  Incremental Tests Running - support for selective tests running based on project model dependencies and the build's changes. This is available for IntelliJ IDEA project runner, Gradle and Maven runners. The functionality is new, so expect issues and make sure we know them;
  • Build Performance Monitor - build-level report to analyze CPU, memory and disk usage on the agent machines.
Also, check out new build failure options - for example you finally can fail a build if it's coverage is below the one of the previous build.
 Check the full list and details on the new additions in the EAP change log.

This EAP updates the data structure, so you will not be able to downgrade to 6.5.x releases after the upgrade. Consider checking EAP features on a test server or prepare to restore 6.5.x server from backup if something goes wrong.

Hope you will take time to look into the features and let us know what you think.

2 comments:

Eric Hauser said...

Any chance we will see an SBT build runner in this release?

Unknown said...

Eric,
SBT support is not yet planned and so far is not on top of our priority list (partly because we do not use Scala ourselves so far and have no experience with SBT). However, there is already a third-party test reporter plugin, so you can already use SBT, invoking it from command line runner. Comments in the issue on what would be the next largest improvement are welcome.