SRECon 2024 happened in Dublin. A few of us went, joined sessions, and talked with folks dealing with the same kinds of challenges we see every day.
The takeaway?
The conference was very much worth attending to. There are definitely some lessons to be learned from it. On the other hand, there are topics that we at Relout are either aligned with, or a step ahead when it comes to the industry as a whole.
Big Themes
A common thread across talks: “Tech is easy, people is hard.” It came up a lot, which absolutely makes sense. Many of the hardest reliability problems aren’t about systems but how teams work, communicate, and and stay on the same page.
What People Were Talking About
- Service Catalogs & IDPs: Backstage came up constantly. Some love it, others are hitting limits. A few vendors pitched simpler alternatives; tools like Port got some attention due to their SaaS model approach and lessening the burden on the maintenance teams.
- SLOs, Revisited: Companies like Zalando shared how they’re rethinking SLOs. Their “WORM” model (Weekly Organizational Reliability Meetings) stood out.
- Weekly Organizational Reliability Meeting (WORM) – an approach Zalando takes to make sure their users operate on a stable platform. The entire concept works on an automatically generated google sheet doc, that pulls data regarding each business area SLO’s, which is then being reviewed on a weekly basis by crucial stakeholders to decide on their focus areas and highlight potential current issues. This is by far the best implementation of SLO’s “process-wise” that Adam has seen up to this point.
- Weekly Organizational Reliability Meeting (WORM) – an approach Zalando takes to make sure their users operate on a stable platform. The entire concept works on an automatically generated google sheet doc, that pulls data regarding each business area SLO’s, which is then being reviewed on a weekly basis by crucial stakeholders to decide on their focus areas and highlight potential current issues. This is by far the best implementation of SLO’s “process-wise” that Adam has seen up to this point.
- AI in SRE: Adam was personally really surprised to see how little AI was being put into a spotlight. From the sessions being presented, to people he’s had a chance to talk with, it seems that the general hype is really tiring for the SRE/DevOps part of the industry. There was only one session touching on AI topic directly (Generative AI: Beyond (Just) Hype by Todd Underwood)
In Between the Talks
As always, the real value often came from chats between sessions – exchanging stories, comparing what’s working (and what’s not). Also: Guinness Storehouse didn’t disappoint.
The Takeaway
Overall, the conference was very much worth attending to (in Adam’s opinion). There are definitely some lessons to be learned from it, like re-evaluating our approach to SLO’s (see WORM at Zalando), or Service Catalogs (Backstage). On the other hand, there are topics that we at Relout are either aligned with, or a step ahead when it comes to the industry as a whole, like for example observability tech stack, working on a engineering platform (k8s) and making sure security is included in the development pipeline.
Anyway, SRECon 2024 was a good reminder that most teams – regardless of size or stack – are dealing with the same stuff. Tools help, but the tough part is still people, processes, and staying aligned as systems grow – and no one knows it better than we do at Relout.
Based on a write-up of the conference by Adam Makowski – Site Reliability Engineering Manager at Relout
