← Back to all sparks
A

AppSignal

INFRA · APIS
Velocity5.0

Application performance monitoring and error tracking for Ruby, Elixir, Node.js, and more

AppSignal is turning distributed tracing into the default way to debug, one link at a time.

distributed-tracingapmobservabilitymcpapi-security
◆Current state
AppSignal shipped a run of tracing upgrades in two weeks: the service map left its feature flag, traces now link to host metrics and deploys, and cross-app services show by name instead of 'Unknown service'. Alongside that, OAuth tokens and the GraphQL API got tighter scoping and a cleaner log query model.
◆Where it's heading
The work points at connecting signals that used to live on separate pages: errors, traces, hosts, deploys and incidents. An MCP tool that lets agents count logs suggests the same connected data is being exposed to AI assistants, not just human dashboards.
◆Prediction
Expect more agent-facing MCP capabilities built on the same trace and log data, likely starting with trace or error queries.

◆Recent moves

  1. 1d ago

    See services from your other apps by name in the error trace map

    Cross-app services now appear by name on an error trace's service map, closing a gap in the tracing push that started when the map went GA a week earlier.

    View source ↗
  2. 6d ago

    OAuth tokens only see what you granted

    OAuth tokens are now held to the organization and apps a user granted in more GraphQL paths, tightening access control as more tools, including agents, call the API.

    View source ↗
  3. 6d ago

    Jump from a trace to its host metrics and deploy

    Trace tags now link straight to host metrics and the matching deploy, another step in stitching AppSignal's separate views into one debugging path.

    View source ↗
  4. 6d ago

    Log severity filters now live in the log query

    A breaking GraphQL change folds severity filters into the log query string, consolidating log filtering on one syntax as the log product matures.

    View source ↗
  5. 8d ago

    Link GitHub issues to incidents

    Incidents can link to several GitHub issues or pull requests via a mention in a note, tying incident work to the code fixing it.

    View source ↗
  6. 8d ago

    Resize the trace timeline and open traces from browser errors

    The service map drops its feature flag and becomes available to every account, with a resizable trace timeline and links from browser errors into traces. This is the anchor release of the current tracing push.

    View source ↗