what's the thing your engineering team keeps relearning the hard way? the latest #observabilityengineering covers new areas incl. debugging #LLMs in prod, #agentic workflows, #telemetry pipelines, and organizational changes that make or break an #observability program. free copy: go.hny.co/4exB7aq

"Most teams are shipping faster than ever, accelerated by AI. They're not learning faster. They're just accumulating risk faster." — @charity.wtf Observability Engineering, 2nd Edition is out today. Nearly 600 pages. Free. go.hny.co/4ebtFCQ #observability #honeycomb

A slow checkout. Three teams asking the same question: app, API, or network? Join @honeycomb.io + Embrace July 14 to learn how to connect frontend sessions to backend traces with #OpenTelemetry, so SRE, mobile, and web teams investigate with shared context. 📍Register here: shorturl.at/R7Qnp

Slow is the new downtime. That line came out of a roundtable at #LDXLondon and it hasn't left us since. Because it's true, and most observability programs aren't built for it. New post on what the scorecard actually needs to look like now www.honeycomb.io/blog/observa...

Honeycomb is at AWS Summit NYC on June 17th, Booth 819. Come see demos of observability with and for AI, bring a real problem, and enter to win a Loops Lab Music Creation System. Book time with us ahead of the day: honeycomb.io/events/aws-summit-nyc #honeycomb #observability #aws #awssummit

@lizthegrey.com just completed our fourth #Graviton retrospective, migrating from #AWS Graviton4 to Graviton5 and running both side by side for 60 days. We also used Canvas to reproduce part of the investigation. Full data plus a public Canvas to explore yourself: www.honeycomb.io/blog/gravito...

Token leaderboards rose and fell at Amazon and Meta in less than three months. Goodhart's Law has never moved so fast. @aparker.io on what the token ROI debate is really missing: honeycomb.io/blog/it-can-only-goodhart-happen #tokenmaxxing #ai #roi

New #o11ycast digs into how Janaki Vivrekar at Amplitude is tackling AI agent building with evals, auto-tagging, and their own agent analytics product. honeycomb.io/resources/podcasts/ep-91-every-failure-becomes-an-eval-with-janaki-vivrekar #honeycomb #observability #aiagent

Day 2 of LDX3 London. We're back. 🐝 Yesterday was full of the kind of conversations you don't get at most conferences. Today we're at it again — booth 510 if you want to find us. Come say hi. Bring a hard question. #honeycomb #observability #ldx3london

Toronto, we're coming for you. 🐝 AWS Summit, June 3rd, Booth 327. Stop by for a demo of observability with AI and for AI. Enter to win a Loops Lab Music Creation System. Talk shop with our team. Book ahead: honeycomb.io/events/aws-summit-toronto #honeycomb #awssummit #observability

what does it feel like when your observability finally matches the complexity of what you've built? you stop checking if things are okay and start knowing. See how Depot got there with Honeycomb. 🐝 → www.honeycomb.io/resources/ca... #observability #honeycomb

iQmetrix set a goal: catch every incident before a customer does. Then they actually hit it. 100% proactive detection. What got them there was treating observability as something worth getting genuinely good at, and having @honeycomb.io made that possible. More → www.honeycomb.io/resources/ca...

Honeycomb and Cloudelligent are taking an already strong partnership to the next level. Cloudelligent's engineers aren't just familiar with Honeycomb. They're in it daily. Read more → www.honeycomb.io/blog/cloudel... #observability #honeycomb #cloudelligent #aws

Nishi Bhonsle at #O11yCon: "Telemetry is not an afterthought. It is part of your feature development." The line between writing code and understanding what it does in production is blurring fast. Engineers who aren't thinking about that convergence already have a gap to close. #honeycomb

@nathenharvey.bsky.social at #O11yCon: AI adoption is nearly universal. Outcomes vary wildly. AI is an amplifier. The goal is not more AI output. It's building an environment where that output can survive the trip to production without burning out the team. #honeycomb #dora #ai

AI agents are transforming software development—teams now mean humans and agents side by side. Real collaboration needs more than markdown: charts, graphs, interactive surfaces. Honeycomb’s updated Canvas, built on @tldraw.com delivers exactly that. #honeycomb #tldraw #observability #ai #canvas

Day 1 of O11yCon is a wrap. 🎉 Five workshops, three and a half hours, a room full of engineers who came to actually learn something. Thanks to Embrace, Focused Labs, AWS, Cribl, and Liatrio for making it happen. Day 2 tomorrow. Main stage. See you there. #o11ycon #honeycomb

Query results, charts, hypotheses. All pinned live. Humans and agents working the same surface on investigations in real time. That's the new Canvas. Available to customers now. www.honeycomb.io/blog/honeyco... #observability #honeycomb #ai

Debugging AI agents in production is brutal. Most tools show you fragments, not the full story. Agent Timeline fixes that. Now in Early Access. www.honeycomb.io/blog/agent-t... #honeycomb #observability #ai

Honeycomb has achieved the AWS Financial Services Competency. 🎉 Built on real production work with teams where #observability isn't just a nice-to-have. Payments, lending, digital banking. High stakes, high cardinality, no room for guesswork. www.honeycomb.io/blog/honeyco...

We believe frontend observability belongs in the same practice as everything else. Today, we're announcing a partnership with @embrace.io so teams get user context from mobile & web in Honeycomb, and powerful, purpose-built RUM in Embrace. bit.ly/48XZBaJ #innovationweek #honeycomb #embrace

Day 2 of Innovation Week is on demand now. Agent Timeline. Canvas. Canvas Agent & Skills. All the details, with real scenarios. Watch → go.hny.co/4clLpug Day 3 tomorrow. Worth showing up for.

Today, Honeycomb is excited to announce: ✅ Agent Timeline: trace every LLM call, tool invocation, and agent handoff as a single view ✅ Canvas: supercharge your team's investigations with an autonomous workspace and expert-encoded playbooks #Observability for the agent era: bit.ly/4wwlzMf

🚨 New #O11yCast! @coryondrejka.bsky.social talks Outcome Engineering: when shipping gets fast, validation is the job. He, @krimple.bsky.social & @jessitron.bsky.social chat agentic AI, the shifting engineer identity, and why measuring real-world impact matters 🎧 → www.honeycomb.io/resources/po...

AI agents can investigate production systems fast, but humans still play a critical role in reasoning through incidents. At #O11yCon, Bryan Mills shares how Duolingo combines AI agents with humans in the loop to improve incident investigations. Register here 👉 buff.ly/o31X0j6

Logs scale faster than understanding. The OpenTelemetry Collector’s drain processor helps cut through log noise by grouping repetitive patterns automatically, making it easier to spot what actually matters at scale. Read more 👉 buff.ly/P9AdWjo

More code is shipping faster than ever, and much of it wasn’t written by a human. That changes how teams debug and build confidence in production systems. See what we’ve been building at Honeycomb Innovation Week. May 12-14 | 10am PT | Virtual Register 👉 buff.ly/osHO1re

AI is moving fast into production, bringing new challenges around trust, risk, and control. At #O11yCon, Giles Douglas shares a CISO’s view on adopting and scaling AI, and how teams are building trust as the landscape evolves. Register now 👉 buff.ly/o31X0j6

AI changes how you investigate production systems. You’re guiding an investigation, not just writing queries. This Honeycomb Academy course shows how to use MCP to ask better questions, structure investigations, and validate your findings. Learn more 👉 buff.ly/f9oY6hS #honeycomb #AI #LLMs

AI agents are reshaping how teams build and debug in production. #O11yCon 2026 brings together the teams working through that shift. Early Bird pricing ends May 11. Get your ticket 👉 buff.ly/o31X0j6

The engineers who built the category are taking the stage. Innovation Week opens with Christine Yen and Charity Majors, alongside teams already running AI agents in production. May 12, 10 a.m. PT. One week out. Register 👉 buff.ly/osHO1re

AI in observability is getting closer to something real. You can see it. You can use it. But it’s not fully there yet. Austin Parker explores what’s working, what’s missing, and what it will take to close the gap. Read more 👉 buff.ly/iv3PkCK

OpenTelemetry semantic conventions evolve, but systems don’t update in sync. During migrations, telemetry gets fragmented across old and new formats. This blog explores how the OpenTelemetry Collector can normalize data in flight to keep observability consistent through change 👉 buff.ly/QdRVK4s

SRE relationships don’t fail all at once. They drift. At #O11yCon, Will Hegedus (Senior Manager SRE at Akamai) breaks down why engagement models stop working and how to fix them before trust erodes. Register now 👉 buff.ly/KOBRNNx

We've been building something. You'll see it first at Innovation Week. New capabilities & live demos. May 12–14, 10 a.m. PT, fully virtual. 👀 Be first 👉 buff.ly/q8uoM0B #honeycomb #observability #AI

AI systems fail in ways that are hard to see. If you can’t trace what your agent is doing step by step, improving it becomes guesswork. Fast feedback loops come from visibility: Instrument with OpenTelemetry → trace behavior → iterate faster. Read more 👉 buff.ly/afUqAU5

AI is changing how we debug production. At #O11yCon, Rob Miles (Head of Observability at Stripe) is sharing what it actually looks like to build an AI agent that investigates incidents in real time. What works, what doesn’t, and what teams should know before building their own. 👉 buff.ly/KOBRNNx

That’s a wrap on AWS Summit London 🇬🇧 AI is reshaping how teams build and run in the cloud, and that showed up in every conversation. Great chats at our booth around real production complexity and how to make sense of it faster. More to come 👉 buff.ly/r1S1e8Z

Production issues don’t stay in one place. They move across services and show up as disconnected signals—hard to piece together fast. The advantage is being able to follow one path through the system, see where it breaks, and why. That’s single-thread visibility 👉 buff.ly/EbnUxIX

ICYMI: When AI writes most of your code, how do you figure out if it’s worth running? The fundamentals haven’t changed: Fast feedback loops Context-rich data Production validation Cost awareness Read more 👉 buff.ly/pcMGnnA

We're doing something new. Innovation Week: Observability for the Agent Era. A 3-day virtual event, May 12-14, 10 a.m. PT. Christine and Charity on stage. New products. Real customers. The agent observability conversation that needs to happen. Register 👉 buff.ly/q8uoM0B

Tomorrow’s the day! We’re heading to AWS Summit London at ExCeL—stop by booth G5 for AI-native observability in action, and don’t miss Colin Burke’s session with Romain Blanc (Intercom). Come say hi, grab a demo 👋 Schedule time 👉 buff.ly/LOiwuni

What if you could cut observability overhead without changing your code? Keep the OpenTelemetry API, swap the SDK → lower CPU + memory. Read more 👉 buff.ly/jo2EAFp

Fast, deep, ready—that’s been the foundation of observability. As systems grow more complex and change speeds up, the real question is what comes next—and how teams evolve to keep up. Read more 👉 buff.ly/WbZZ1Em

We’re excited for all our speakers at #O11yCon, May 20–21! Jody Bailey (Stack Overflow) will explore how AI agents are changing how we think about organizational knowledge in production—and why gaps in context are becoming real risks. Don’t miss it 👉 buff.ly/2FQ3IWX

We recently hosted a webinar with Austin Parker and Akshay (Augment Code), and you all brought some great questions—more than we could answer live. We followed up with a blog covering topics like trust, drift, and production feedback for AI agents. Read more 👉 buff.ly/2zSTpKP

AI agents are starting to work directly with production data, investigating issues, following traces, and surfacing what’s happening in real systems. This post explores how MCP and Canvas support that workflow by connecting context and action. Read more 👉 buff.ly/zX7nqF1

#O11yCon 2026 is coming. May 20-21, San Francisco. 🐝 AI agents are running in production. Can you actually see what they’re doing? 2 days of workshops + sessions with speakers from Google, Stripe, GitHub & more. Early bird tickets are $99 until April 28. More 👉 buff.ly/N89JR5k

If you’re attending AWS Summit London, be sure to stop by Colin Burke and Romain Blanc’s session! Learn how Honeycomb + Intercom introduced a customer‑centric “time to first token” metric, rebuilt distributed tracing, and used SLOs to cut latency by 60% while 20 teams ship ~100 changes/day.

Uncertainty and change are the default in modern software. As systems grow more complex (and AI speeds things up), rigid plans don’t hold. Success comes from fast feedback loops, learning, and iteration. Read more 👉 buff.ly/L2M34c2