Today I joined a DORA community session with a talk about an open source project, “CDEvents”. I’d not heard of it before and it sounds super useful for organizations large enough that different teams may be using different deployment tools etc. I’e seen this myself - there’s one product, which consists of subproducts built by different teams. One team uses GitHub Actions, one team uses Jenkins… Among other problems, say the product is undergoing big changes and you want to see how they affect all the automated tests for the whole product. You have to search around in multiple places for test results.
The CD Foundation is building a common specification for all continuous delivery events. A standard event collector across all tools used in CI/CD would let you put all test results from all platforms in one place. Of course it also allows collecting all CD events and info including DORA key metrics. The goal is to span the entire SDLC. It sounds like such a good idea and one where I think, why didn’t someone think of this before?
I’m curious if anyone here is in this type of larger scale situation? Does this sound like a useful thing, having this common “language” of continuous delivery workflows?
So that could be due to tech sprawl within the company or different needs for different teams but in general they all should have APIs or someway to report to a central hub for things like DORA Metrics. You could use an (internal developer portal) IDP or monitoring tools (splunk/dynatrace) to capture all these and report them back.
Each tool or deployment pipeline isn’t always a one size fits all. We have Argo, Jenkins, Github, harness, rollout and that’s only a couple of the tech stake we have because we need them for different technologies, however they all report out to a central monitor that we all have access and can view, dynatrace for deep details, and our IDP for those who want high level views.
So while yes, having everything in one spot can be really great and easy sometimes depending on the tech stack that may not be feasible, but the reporting that could be sent to a single reporting structure.
It sounds like there could be some tech sprawl in your area which would make it difficult to track it all down but (with super limited knowledge and a lot of assumption on my part) sounds like you might need a better observability and monitoring platform/practices to aggregate and gain insight to these areas easier.
I don’t have this situation myself, I’m a freelance consultant. I was just impressed with this open source initiative to create a common specification that everyone can use, it’s tool-agnostic. I can’t explain it well, this is a good explanation. CDEvents addresses this core challenge by providing a mechanism by which you can integrate all the legacy components of your build processes into a unified Continuous Delivery process for your Product. By adding a layer of events, sharing a common vocabulary, to each phase of all of your build processes, you can integrate all these stages into a coherent process providing a single view of the state of all your assets.
I’m sure it can be done other ways, but this could reduce a lot of toil and provide an industry standard. The other leading CD practitioners in the meetup seemed quite enthusiastic about it and ready to contribute.
On the face of it, it sounds like it could be really useful for organisations in this situation. Upon reading the comments, I started to wonder whether the same thing might be achievable using third party aggregator / dashboard tools. I’m not super into the topic though, and haven’t been in this situation, as far as I can remember, so I can’t say much else apart from that.