Jepsen tested Orkes Conductor, a workflow orchestration platform that runs user-defined tasks and routes data between them, across versions 5.3.2 through 5.5.0-rc2. The analysis found unexpectedly frequent concurrent execution of individual tasks, which could significantly delay workflow execution, along with four minor start Workflow errors or null IDs.
The most serious findings were three safety problems. In these cases, Conductor permanently abandoned or forgot workflow executions, or marked them completed even though their tasks were never run. Orkes has released version 5.5.6 with patches for all serious safety issues, reduced concurrent execution, and fixed two of the start Workflow errors. The remaining two are under investigation. Delayed execution is consistent with Conductor's design parameters and need not be addressed.
Conductor is a system originally built by Netflix and open sourced in 2016. Its open source version is now maintained by the Conductor OSS Foundation, which is led by Orkes. Orkes also offers the hosted Orkes Conductor, which is the focus of this analysis.
The research was funded by Orkes Inc. and conducted in accordance with the Jepsen ethics policy. Orkes has published a companion blog post to this work.
Source: Hacker News · Summarized by HeadlinesBriefing