LoginSign Up
← Back to all sparks
A

Axiom

ANALYTICS
Velocity5.0

Log management and observability platform for events and traces.

axiom.co ↗

Axiom makes agent work checkable, and points alerts straight at coding agents

observabilityagent-nativemcptracingalertinglabels
◆Current state
Axiom's recent releases split between agent plumbing and trace analysis. Monitors can alert Cursor, Claude Code, Amp and ChatGPT, MCP links open the exact saved result an agent returned, and Critical Path, Service Map and Statistics views arrived for traces in October. The 6 November release added a phone-ready console and a new sidebar, and a 12 November post confirms every page now fits phones, foldables and tablets.
◆Where it's heading
The agent track has moved from access to accountability. Saved result links and Open in Axiom links exist so a person can inspect what an agent saw and built, and notifier delivery history applies the same scrutiny to alerts. Sending monitor alerts directly to coding agents closes a loop from detection to attempted fix without a human relay. The trace work is a reminder that the product is still being built for human investigators as well.
◆Prediction
Expect further notifier types and agent skills. Whether alerts sent to coding agents become a measured remediation workflow is unclear from these entries, which describe delivery rather than outcomes.

◆Recent moves

  1. 2d ago

    Console for phones, foldables, and tablets

    A short note restating the mobile console that shipped on 6 November: dashboards, charts and traces fit phones, foldables and tablets. No new capability beyond that release.

    View source ↗
  2. 8d ago

    Console on your phone, a new sidebar, and slow spans at a glance

    The console now works on a phone, navigation moves to a new sidebar, slow spans are flagged at a glance, and monitors can alert Jira Service Management. It extends October's trace views and alert targets rather than opening a new direction.

    View source ↗
  3. 16d ago

    Agent work you can share, notifier health, and Amp and ChatGPT alerts

    Agents' work becomes shareable through links to exact saved results, and two skills let agents send data, including into a temporary organization. Notifier delivery history and Amp and ChatGPT alerts round out the release, with trust in agent output as the common thread.

    View source ↗
  4. 17d ago

    Service Map and Statistics for Traces

    Service Map and Statistics views summarize large traces, including uninstrumented databases and hosts seen through client spans. They follow Critical Path as a run of trace analysis work aimed at human investigators.

    View source ↗
  5. 23d ago

    Critical Path for Traces

    Critical Path highlights the spans a trace actually waited on, using an approach based on Uber's CRISP algorithm. It is a focused way to find what to speed up in traces with thousands of spans.

    View source ↗
  6. 24d ago

    Monitor alerts for coding agents, one-paste MCP setup, and search inside traces

    Monitors can now hand alerts to Cursor and Claude Code, and a copied prompt connects a coding agent to Axiom MCP with OAuth and no token. This extends agent-native work from reading data to receiving incidents.

    View source ↗