Fred Hebert's team at Honeycomb deliberately leaned into the code review bottleneck caused by AI-generated code, choosing to maintain thorough reviews rather than reduce them. The team found that velo…
Honeycomb processes 2 million agent-initiated query runs per month, up from a $0.60 eight-tool-call demo a year ago, with average tool calls per session doubling between February and June, according t…
AI agent failures in production stem from poor architecture—unmanaged context windows, missing governance layers, and insufficient observability—rather than model capability gaps, according to a Hacke…
Honeycomb engineering team 2.5x-ed their throughput using AI without breaking quality, but the key lesson is that AI amplifies existing practices—it cannot fix a dysfunctional organization. The team e…
Honeycomb's engineering team more than doubled peak-weekday merges from ~30 to ~74 by April 2026, with AI-attributed code rising from near-zero to 82.6% of new code by June 2026, according to a blog p…
ClickHouse's 26.6 release introduces hypothetical skip indexes, AI embedding functions, and experimental continuous queries, while a new case study shows Visa's Authorize.net team using ClickHouse Clo…
Honeycomb completed a multi-month migration of its Kafka infrastructure from self-hosted Confluent Platform and ZooKeeper on AWS EC2 to open-source Apache Kafka 4.1.1 running in KRaft mode on AWS EKS.…
Fin CTO Darragh Curran writes that AI is an amplifier of existing engineering practices, not a magic wand, and that engineering rigor remains essential. The second edition of Observability Engineering…
Monitoring catches the failures you predicted, but AI systems fail in ways you didn't, according to Kale Bogdanovs in a post on the Honeycomb blog. The post argues that AI workloads demand observabili…
AWS Graviton5-powered M9g and M9gd EC2 instances are generally available, with production benchmarks from Honeycomb, ClickHouse, and HubSpot showing consistent 36% performance gains over M8g. The new …
Honeycomb built a multi-agent collaboration architecture for its Canvas investigation tool, where multiple engineers' AI agents operate independently in the same runtime session while sharing awarenes…
Honeycomb has shipped a production integration with Amazon Bedrock AgentCore, surfacing agent telemetry in its Agent Timeline tool. The company chose AgentCore for its modularity and production-grade …
AWS made Graviton5-powered M9g and M9gd instances generally available on June 10, featuring 192 ARM cores and the first formally verified VM isolation in commercial cloud history. The new instances de…
Honeycomb's Charity Majors defends AI mandates as a necessary tool for funding and executing coordinated change on tight timelines, arguing that without mandates, companies signal AI is not a priority…
Reid Savage, an engineering manager at Honeycomb, reflects on his first year leading the SRE team, emphasizing the importance of investing in relationships, seeking feedback, and distinguishing betwee…
At O11yCon, engineering teams from Mixpanel, Gem, and StarSling reported using AI to boost code velocity by up to 50%, with observability tools like Honeycomb critical for managing the increased volum…
Honeycomb released the second edition of Observability Engineering, a near-complete rewrite addressing AI-era challenges like validating AI-generated code in production and the risk of shipping faster…
Honeycomb released a technical guide on instrumenting AI agents with OpenTelemetry's GenAI semantic conventions to enable debugging via the Agent Timeline, which captures tool calls, multi-agent hando…
Honeycomb's engineering team redesigned their parallel test execution system to prevent database collisions when multiple pytest invocations run concurrently against the same PostgreSQL host. The fix …
A developer advocates instrumenting AI agent decision tracing with OpenTelemetry to enable rapid incident response. The approach uses spans to capture reasoning, context, and tool executions, making a…