Cavaillé Léo (@leoc892) profile picture

Director, SRE @datadoghq. Also scuba diving, woodworking, electronics & eating/cooking things.

Paris, France
Joined October 2010
Photo by Cavaillé Léo (@leoc892)Photo by Cavaillé Léo (@leoc892)Photo by Cavaillé Léo (@leoc892)Photo by Cavaillé Léo (@leoc892)Photo by Cavaillé Léo (@leoc892)Photo by Cavaillé Léo (@leoc892)Photo by Cavaillé Léo (@leoc892)Photo by Cavaillé Léo (@leoc892)Photo by Cavaillé Léo (@leoc892)Photo by Cavaillé Léo (@leoc892)
"which ultimately led to a series of unfortunate events"
1
Cavaillé Léo retweeted
Two awesome folks from my team (Tay Nishimura and Joris Bonnefoy) will be talking about Chaos Engineering at Datadog tomorrow June 1st at 6PM CEST ! you can get in the event for free here : datadogon.datadoghq.com/talk…
5
1
16
Every time I am AFG (Away From Grinder) for a couple of days, I forget this freshly ground coffee smell makes my morning routine so much more enjoyable 😋☕️
After this long period without logging any time under water...! Chilly but fun 🐟😎🥶
3
How you can easily catch my attention 😁... "I actually enjoy incidents as I always feel like this is an opportunity to learn"
1
4
Days since I left a DNSSEC zone signature expire = 0
1
"Reliability is not a team". That's mostly the tweet. 1/2
1
5
It can be startling for some folks when I say that and my job is actually... to manage and spin up new SRE teams. I insist :), SRE's mission is making sure to *not* keep reliability and production expertise and experience inside the SRE group itself. 2/2
2
Corollary: If you can't measure it, you can't say it's fixed :-)
1
3
Realizing today I haven't been able to publish anything in 10 months, not even nerd jokes or food. These last 10 mo definitely had the highest of highs and lowest of lows... Well, going to try to stop doomscrolling and share some cool ideas/stuff with you to break the cycle ☺️
1
7
Vote for me and I will sign an executive order forbidding CAs to sign certificates which will expire on long weekends.
1
3
1
26
If a fix was deployed for an incident but the team is not able to tell when the same factors & conditions that led to the incident are happening again, nothing is fixed actually. Visibility and awareness 👑.
1
This👇will be printed in the header of every game day 😬. Grateful for people taking the time to go deep and write more about this topic. Usually the revelation comes with the chaos experiment where a team learns about itself rather than a bug.
Replying to @nora_js
Chaos Engineering is not about finding or surfacing vulnerabilities.
2
8
Seeing hot takes from the community on the "5 whys" framework pitfalls for incidents analysis so decided to share mine: 1. Why? 2. Wh..ow wait a minute, is this possible? 3. Wh..oof no, thank god 4. Wh..TF scratch that, it did happen 5. Wh..e misinterpreted signals
5
To me from 2016: "tracking ongoing actions and hacks to undo is essential during incidents". Someone just hit me up yesterday with: "I ran into this key -> 'OUTAGE_QUEUE_20160926_LEO', is this okay to delete?"
5
Do I really want to add SPOF to my dictionary...
1
3
If you ask me about the cost of having backups I will always return you the question of what is the cost of not having them when you'll need them ;-)
5
Just got back on track after being sick and was about to enjoy a few days off even, damn it dryer
1
5
Found the SPOF ugh, high availability is a lie
1
Distributed systems give you distributed bugs
2