2 Days are hardly enough in a city like Barcelona – especially when you spend most of the time in an office and don’t get to enjoy the vibrant life of a city on the mediterranean.
But – my last trip to Barcelona early May – while it was short – was more than I could have asked for. And here is why!
Meeting our users and having them listen to me
In my job I am fortunate enough to have people listen to me – even when I don’t speak their language. And most of the time I think they are honest when they tell me that it was time worth spent 😊
While in Barcelona I was asked to spend an hour with a team of a large bank that are on the journey to adopt Dynatrace. My mission – as always – was to discuss how to really get value out of the investment in observability tools like Dynatrace. To keep it short. My key message typically is:
- Bring Insights TO the developers, INTO the tools where they do their work
- DO NOT ask them to become an expert in yet another discipline or tool
The following is a screenshot where I give 3 examples of how this can look like:
Bringing observability into the IDP, Pull Request and the IDE / Coding Agent
My examples from above are typically followed by demos where I show:
- Backstage (IDP): Make relevant logs, metrics or vulnerabilities accessible
- Git (Pull Request): Provide an automated observability score for every code change
- VSCode (IDE): Enable the coding agent with an Observability Skill and MCP
I wanted to say “muchas gracias” for the time that those teams are spending with me! I am looking forward seeing some of my examples being implemented.
Now – if you are interested in seeing some of those demos then check out my postings on My LinkedIn!
DevOps BCN Meetup
Thursday evening, we hosted DevOpsBNC Meetup in the Dynatrace office. I was the second speaker with a very AI-focused topic. To my surprise – or maybe to my relief – the first speaker started his talk with a promise to not talk about AI. And he fulfilled that promise! More on the “not everything needs AI” later.
Don’t always follow the herd!
The talk started with the shift from monoliths to microservices and which problems it solved but which new challenges it brought. For me the key message and a warning was: “Do not always follow the herd!” The speaker brought great examples on how technology teams sometimes blindly follow everything that the “the crème de la crème” organizations do. Like blindly following Googles SRE practices or Netflix’s way to build microservice architectures. There are countless examples of tech giants (and unicorns) that the tech industry worships and often blindly copy/pasting their approach of adopting technology.
What I often miss is the courage to discuss and challenge the blind adoption of so called “best practices” just because they worked for one organization. I also sometimes miss the constructive conversations as we are too easily convinced by one approach and then defend it without good reflection.
Observability in the AI-Native Age
My talk at the DevOps BCN meetup was titled the same way as our recently released book: “Observability in the AI-Native Age”. In my talk I explained the role of Observability to observe AIs but also to provide closed loop feedback to the AIs to continuously improve its output as we use it in the AI Delivery Lifecycle. Below my opening slide that highlighted my talk track:
Observability plays a critical role in the AI-native era we are just entering
And while AI is here to stay I also added my thoughts around that not every problem needs AI to solve it. I typically bring two examples where simple automation can provide enormous value to engineers as you can see here:
An automated daily message and an automated score on a Pull Request are powerful feedback loops!
Instead of prompting an agent for insights, valuable insights can easily be pushed to the engineers, into the tools they are using, at the right time when they need it, e.g: a Slack message 5 minutes before the daily standup that shows findings in observability data!
I wanted to say Thank You to everyone who showed up to that DevOpsBCN Meetup. And for everyone in Barcelona – do not forget that DevOpsDays Barcelona is coming up in November and that there is still a chance to submit a talk proposal: LINK HERE!!
SREDay: Stranger Things, AI, Save a Tree and DevOps In Space
Friday was SREDay Barcelona, the first of its kind in the Catalan capital. It brought together a diverse set of voices across SRE, DevOps, platform engineering and cloud infrastructure. From discussions on observability and resiliency to deep dives into Kubernetes networking and chaos engineering, the event highlighted one clear trend: modern reliability engineering is increasingly shaped by automation, AI, and proactive experimentation.
My Stranger Things inspired Platform Engineering talk
I was fortunate to kick off the event with my talk “Stranger Platforms: The Two Sides of Platform Engineering!”. In my talk I highlighted how Observability is connecting the two worlds in a modern software development platform: The Developers (the users of a platform) and The Platform Engineers (those that build and operate the platform).
Observability has the power to connect and benefit both sides of a platform: developers and platform engineers.
I brought a lot of my examples and some live demos where I showed what questions observability can answer for the users and the producers of a platform. The following slide gives a good top (right side up) to bottom (upside down) overview!
Observability provides critical insights for users and producers of a platform and helps making platforms successful
For anyone interested in my slides: feel free to drop a comment here and I am happy to share them as they also include lots of my demo videos and technical advice!
From AI via Saving Trees to DevOps in Space
As a speaker I don’t often have the luxury to stay the whole event and listen to all other talks. But SREDay in Barcelona was an exception. I was able to attend each talk and can attest that every single speaker had an amazing message for the audience! I do believe that all presentations were recorded and they will be made available on the SREDay YouTube channel.
Some of my key takeaways from those talks
- AI SRE Agents are here and provide value by orchestrating the tasks when you have to react to incidents. Whether its collecting more root cause information or preparing a Pull Request with a fix
- Performance Engineering is more important than ever. Without optimizing our systems – especially all those AI workloads – we will pay a high price. The motto of one of the talks was: Save a Tree!!
- If you are providing a multi-tenant SaaS system a simple tip is to create a “Canary Tenant” on your cluster and leverage this to test and validate the end-2-end experience of your “real tenants”
- Explaining networking in Kubernetes can be very edu-taining! One of the most technical talks was about a strange networking problem but the speaker did a phenomenal job for everyone to follow a very deep dive technical explanation!
- DevOps in Space is “out of this world”! The closing keynote was a reminder that space – the final frontier – plays by its own rules. We learned about all the constraints in updating software in satellites that only have a connectivity multiple times a day for a couple of minutes with very limited bandwidth
- And so many other learnings across chaos engineering, resilient architectures or cloud costs
Thanks to the organizers of SREDay for organizing such a great event. For everyone interested in the topic be reminded that SREDay is a worldwide event series. Check out their website and find the closest event to you!
This was my recap of my time in Barcelona. After Barcelona I went on to Toronto for Dynatrace Innovate Roadshow and KCD Toronto. Stay tuned for another trip report!
