Enterprise observability. Less drama, more clarity. Author. Metrics & Mayhem podcast + newsletter.
UK
- Tweets 597
- Following 158
- Followers 107
- Likes 101
Observability Digest, August 2026: The Foundation Nobody Funds masteringobservability.com/p…
14
The Signal: Own Your Telemetry, Rent the Observability Platform masteringobservability.com/p… #observability
21
The Gate Stays Human: Who's Accountable When AI-Ops Acts masteringobservability.com/p…
15
Good tools answer the question so fast you forget you were the one who asked it. A dashboard, a runbook, an AI incident summary. Each is useful. Each retires a question you used to ask yourself.
1
1
15
This is not anti-AI. Automation, mature tooling and plain seniority all do the same thing. They retire your questions. The skill is noticing which questions you have quietly stopped asking.
1
11
The fix is a 20-minute habit: put the reps back on purpose. Take one recent incident and investigate it from scratch, as if the summary did not exist.
Full episode: piped.video/HmUPwluP-Ug
#Observability #SRE #IncidentManagement #AIOps
6
New Signal Drop is live. Good tools answer so fast you forget you were the one asking. Dashboards and AI summaries quietly retire the questions you used to ask. Curiosity is a muscle. Put the reps back on purpose.
🎧 open.spotify.com/episode/68x…
#Observability #SRE #AIOps
15
New Signal Drop: I watched one of the best incident teams I've worked with get quietly worse over six weeks. Every change we made was, on paper, an improvement. The reps just disappeared.
🎧 masteringobservability.com/m…
17
Every observability vendor bolted an AI assistant onto its dashboard. Each is wired to its own data with its own custom bridge, so none see your whole stack.
MCP is the standard trying to fix that. The first Tech Tuesday:
piped.video/xBHxBNhPf5M
16
I once watched one of the best incident teams I've worked with get quietly worse. Over about six weeks. Every change we made was, on paper, an improvement.
1
11
I was watching the wrong number: how well we handled the incidents we'd automated. The number that mattered was how well we handled the ones we couldn't. That was quietly falling the whole time.
1
3
Automate the triage, and you stop training the triager. The fix isn't to stop automating. It's to put the reps back on purpose: one live drill a month, a novel fault, no runbook, your best people, cold. Full episode: masteringobservability.com/m…
4
New Signal Drop: I watched one of the best incident teams I've worked with get quietly worse over six weeks. Every change we made was, on paper, an improvement. The reps just disappeared.
🎧 masteringobservability.com/m…
7
Spent yesterday at Dynatrace Innovate London as an independent voice, not buying and not selling. The vendor demos were fine. The customer stories were the education. Three patterns ran through all of them.
2
1
39
From a talk that had nothing to do with observability: 'The winner of the communication race is the person who speaks last.' A technical leader's real second job is director of translation.
1
1
14