Earlier this week, CloudBees proudly announced the release of ElectricCommander 3.10 – including some really exciting new functionality in the product with the marquee features being focused around more dynamic, efficient, reliable and exact resource usage. With this release, ElectricCommander takes yet another big leap towards being the only true platform flexible and reliable enough to fully support the complex and demanding needs of any organization looking to enable a scalable, secure and self-service development cloud .
With ElectricCommander, DevOps teams are now able to take complete control over how their underlying IT infrastructure are being selected and scheduled to power the development workflows - no matter if this infrastructure is made up of physical, virtual, cloud-based or custom embedded hardware resources. By allowing for configuration and customization of the ordering algorithm within a resource pool, a whole set of interesting use-cases will now be satisfied out-of-the-box by ElectricCommander. Let’s look at a couple of these real-life use-cases, both of which we know some of our customers have been requesting and waiting for:
1. True round-robin resource assignment. In a large scale environment where you have thousands of integration jobs running per day at an extreme speed, it can be of crucial importance for capacity and maintenance planning to know that the workload distributes evenly across your infrastructure (and not randomly based on availability). One customer of ours has an environment that generates loads of data per integration job, data that needs to be stored and managed offline for auditing reasons. By true round-robin resource assignment, the management of the data on these resources will be much simplified, allowing for the Operations team to trigger maintenance and cleanup processes at will with control, without interfering with the integration builds of the engineering team.
2. Least used resource assignment. The speed and throughput of many development processes are bound by certain hardware limitations, such as CPU-load, network bandwidth or disk I/O. To optimize the utilization of their resources in the infrastructure, many of our customers have asked us for ways to select the most appropriate resource out of the pool based on some criteria of physical load – a use case that’s now satisfied in ElectricCommander by the configurable ordering algorithm of a resource pool.
Another key and important new feature in the 3.10 relase is the ability for ElectricCommander to react and block on failed workspaces. This means that whenever a workspace e.g. runs out of disk space or has a failed network connection, ElectricCommander will automatically notify system administrators while trying to keep running jobs alive through waiting and retrying. This is a key feature in the area of reliability that many of our enterprise customers has been asking for - no more failed jobs due to misconfigured workspaces or failing disk drives!
To ease system administration of an ElectricCommander deployment, the 3.10 release is also introducing a system event visualization feature. System administrators and DevOps teams are now able to very quickly monitor the status of the environment with a new interface displaying the latest log of events generated anywhere in the system including jobs, workflows, resources, workspaces and notifiers. Unlike prior to the 3.10 release where all of this log data were available in log-files in the installation and data directories, it is now securely exposed directly within the web interface of the application.
As with any new CloudBees release, there are also a number of minor features and performance enhancements – I encourage you to read through the release announcement for more details and let us know your feedback !