4
2 Comments

Show IH: elrond.io - DevOps collaboration chatbot

Hi Hackers!

I wanted to run a service by you that I'm building. It's called http://elrond.io and it helps DevOps teams collaborate on issues and outages.

DevOps teams often work with chatrooms to communicate and collaborate. Commonly, the tools these teams work with (for development,
deployment, monitoring, etc.) all connect to the chatroom as well to alert the team of one of the systems they are responsible for experiences
an outage.

Whenever an issue occurs, the members on the team self-organise to find someone (or someones) who can effectively troubleshoot the issue. As
long as the team is small, this works fine in a single chatroom with all team members. But once you work with multiple teams managing multiple
applications and environments, this breaks down. It becomes hard to locate the right engineers that know this application or environment,
that fixed a similar issue before or are available to help. Even a simple thing as using a round-robin scheme to assign this work across
distributed team members is challenging.

The Elrond chatbot monitors your chatroom to detect outages in your Continuous Delivery pipeline. When an incident occurs, Elrond looks for
engineers who can help -- engineers that built the software or servers, that introduced the most recent change or that resolved a similar
issue before. Based on factors like availability and timezone, Elrond involves one or more engineers so the issue at hand can be dealt with
quickly and efficiently.

I'd love to hear any feedback you might have and will of course reply to any questions.

Thanks for reading!

  1. 1

    I'm not too familiar with chat bots and how they work, but I do think the idea of automatically connecting a devOps issue to a developer is interesting. Specifically the part about connecting the developer who made the most recent changes (git-blame or similar?). I would think you could try to sell this by saying (if it really does, of course) how it reduces down time and availability and resolves issues faster.

    I wouldn't use this myself as I'm not in devOps, but I do like the idea.

    1. 1

      Figuring out who made the most recent change requires some magic and I haven't yet fully figured out how to do it. On the dev side, one way is to receive the info from the chatroom itself (posted in the chatroom, for instance as output from a bot connected to eg JIRA) but there are others, too.

      For Ops, perhaps it could indeed be taken from git (infrastructure-as-code).

      I'm hoping that the ability to find the right engineer to solve an issue based on software/environment & availability is enough of an MVP. Then I can add the "latest change" functionality afterwards.

      Thanks for your comment!