 Good morning, so my name is Ryan Brady. I work at Red Hat and I've been a user of and a contributor to Mistral for about a year So Mistral is a workflow service It provides you with a mechanism to define tasks and workflows that Manages the the state of those workflows while you're executing in your open stack cloud Mistral was originally founded in the Havana release of open stack Our latest release has about 50 contributors total the recent survey adoption numbers Or that were deployed in 5% of clouds But another interesting note is that we've been in triple O used in all triple O deployment since Newton Going back to take a look at what we did in Okada. There are some performance and reliability improvements There's a lot of bug fixing that went on There are some improvements to the the horizon plug-in that we use is our Mistral dashboard there were Yakkel functions that were added most notably was the tasks function which You can filter all the users tasks by a workflow execution id or state and you can recursively get the entire list of tasks that you need Looking forward towards pike What we're working on right now For workflow error analysis today when you're When a workflow fails it may be difficult for you to quickly find the root cause there for each task that you want to look at and Inspect you have to run three different commands in the client And you'll have to keep doing that over and over to get all of that data back And then you compile it yourself to take a look at it So this feature is where we're gonna have a single API call It's gonna go get all of that information for you and return that back to you in a more usable way So it should cut down your time of trying to get to the root cause of whatever the failure is For the GUI This is going back to the Mistral dashboard There's gonna be a demo on the 25th and an email is gonna be sent out to the list But the the dashboard is crucially important, especially if you're dealing with Big workflows or long running workflows And it's hard to see the whole picture clearly. So we want to provide some sort of visualization to you and then We're refactoring our open stack actions There's a larger we'll talk about that later, but there's a larger Change to our entire action Platform throughout and so we need to refactor open stack actions on top of that but we also want to Provide some multi versioning support. So if you wanted to use You wouldn't be pinned right now right now if you wanted to use say ironic You'd be pinned to whatever release we're using for the ironic client But we want to be able to provide you the ability to pick which client you wanted to use if you're using a different version So the multi region support to allow you to run different actions in different clouds or different tasks in different clouds It's gonna allow it to run without any Configured keystone in point I'll actually be performed using whatever keystone in point credentials are passed into the request For the sensitive data if you want to pass like say passwords into a mr. Workflow right now and it gets logged somewhere that password could Potentially be logged. So the sensitive data spec is Going through and ensuring that you can declare what data needs to be kept private in a workflow That you're passing in and then that'll be masked in areas like logging so that you won't be exposing some of your secrets The workflow execution event publication and subscription Right now you can interact with Mistral from an external system You can create workflows and workflow definitions. You can schedule them to run you can check their execution But to do all this you're gonna end up polling to get the execution and to to get what the status is and then Come back and get the result so this feature is essentially creating a Callback API if you will so you can pass your call back in whenever the workflow completes It will call back to the other system and it's gonna be implemented via plug-in. So you would create a Plug-in that would allow you to do this the release themes So scalability interoperability user experience are all things that we're really tightly focused on a little bit on manageability and security Modularity and resiliency aren't something that we're gonna be focusing on in pike, but heading into the future We will be looking more at some of those So instead of breaking out into Queens and the our release We've got some project goals that we published for 2017 and these are generally gonna drive What we're doing as time goes on so the first one is our documentation. We really need to Close the gap on things that are missing we need to have a consistent theme in our docs and a consistent style kind of a Overall improvement we need to ensure that all the examples and tutorials are published over in Mistral extra that they're working That they're up to date We need to complete our custom access API for two cycles We've been working on revamping How actions work for Third-party support if you want to integrate, but it's causing a big change internally to change all of our actions to follow that We also want to work on multi node mode And there's a lot of work that is involved with making sure that that runs smoothly and The last one is we really want to increase performance We want to reduce average workflow execution time by 30% and that's something we feel we can definitely do So if you'd like to find us we're always looking for contributors We hang out in OpenStack Mistral on free node and we have weekly meetings at 1500 UTC and OpenStack meeting three are there any questions? Yes, sir Your question was what were the primary underpinnings? That would change that would cause increases in performance. Okay. Good question We're going to be working on part of the RPC work that's underneath There's improvements to the database layer that we can do Pretty much that's what I know for now We can talk to her not. Yes, sir. It's a very interesting point. I think I'll definitely take that back to them Thank you. Yes, sir your question will So his question was is Mistral a good use case for cleaning up VMs that may have been Running too long. Was that your question? Okay. So yes, Mistral seems like it would be a great choice you could schedule your workflow with Cron trigger and that workflow could go out and check the status of it and Take care of whatever you need to there's OpenStack actions that where you could integrate with Nova there Send it to your question Awesome, I'll talk to the triple-O guy that works on Mistral. Are there any other questions? Thank you for coming out this morning, and I hope you can find us an IRC if you guys want to come Learn about Mistral or get it up and running or try writing a workflow