Maintaining the infrastructure and operational systems for optimization models requires more than wrapping up or containerizing a model. From creating observable systems to managing the load of large-scale simulations, there’s a lot of engineering coordination and effort involved.
At Nextmv, we’re empowering decision and data scientists to focus on building models, and freeing up engineering teams to work more efficiently with a service that fits seamlessly into their tech stack. Queueing and prioritization on Nextmv is one piece of that puzzle. Modelers can initiate runs with the peace of mind that their high-priority jobs will run first and any long-running jobs will not cause a large backlog. There’s no need to write custom scripts as workarounds, request engineering teams to configure and set options in your infrastructure, or cross your fingers and hope for the best. Let’s take a look at how simple it is to get started.
Prioritize optimization model runs
Whether you’re submitting large batches of concurrent runs or need to manage a mix of production and staging runs, adding a priority ensures that your queue is ordered strategically and keeps operations running smoothly. Simply choose between 1 (highest) and 9 (lowest) to rank the run’s priority via the UI, API, or CLI. There are two ways to set it: at runtime or at the instance level.
Configure a priority at runtime. You may want to set priority at the run level to give shorter runs precedence, so long runs don’t block the queue. Below, we’ve set a run with a queue priority of 6.
Configure a priority per instance. You may want to set priority at the instance level so all of your production runs take precedence over any staging or experimentation runs. Below, we’ve configured the queue priority on our staging instance.
View and manage run queues
Once you’ve submitted your runs, you can view them all in the Nextmv UI. Use this dashboard to spotcheck priority, view the status of each run, and click into run details to see configuration, input, and more. Below, we can see runs in the queue ranked by priority along with their status, create date, and other metadata.
Get started with queueing
Ready to simplify your optimization model management with queueing? Sign up for a free Nextmv account to get started. Have questions? Reach out to our team, and we’ll be happy to help!
