Dash0

Recently shipped AWS Lambda observability with one-click instrumentation and Agent0 running autonomously on event-driven triggers.

Agent0 can now be delegated work through the Dash0 MCP server. Connected assistants like Claude query metrics, logs, traces, and dashboards and can start investigations against your live environment. Agent0 works out which queries to run and reports back within a conversation. Read-only only — outputs are dashboards and alerts proposed for your review.

Full AWS Lambda observability is now GA. Connect an AWS account via CloudFormation and see every function immediately with traces, correlated logs, and synthetic traces for failures. One-click instrumentation from the UI works for Python, Node.js, and Java without code changes, with automatic drift detection. Also includes payload capture of Lambda events and downstream AWS service calls with sensitive value masking.

Spam filters gained metrics showing how much data they drop. Each filter displays a time series of dropped signals (logs, spans, metric data points, web events). These metrics are automatically recorded and don't count toward billing — you can also alert on trends using check rules.

UI improvements:

• The span sidebar's events tab was redesigned to "Related logs & events" with an interleaved, time-ordered list of log records and span events across trace, resource, or both. Includes inline marker rows for span boundaries and deep links from the trace tree.
• The new Spans Table widget brings a sortable, self-refreshing table of raw spans to dashboards. Each row links directly into the Trace Explorer with columns and filters preserved.
• Any query in the Log, Trace, or Web Events Explorer can now be exported to CSV, TXT, JSON, or YAML. Supports up to 5,000 records; for larger exports, use the Dash0 CLI.

Teams are now manageable as infrastructure as code using the Dash0 CLI (v1.16.0+), Terraform provider (v1.15.0+), and Kubernetes operator (v1.50.0). Team definitions live in the same repository, review process, and rollout pipeline as everything else.

Operator updates across recent releases:

• Installations on GKE AutoPilot should remain on Operator 0.146.0 — releases 0.147.0 and later are not compatible.
• Signal Control (formerly Intelligent Edge) restructured pipelines to support namespaced and multiple datasets. "Barker" was renamed to Edge Proxy.
• Token redaction in logs: the Dash0 self-monitoring token was visible in operator logs, now redacted. Consider rotating tokens.
• Python auto-instrumentation now requires Python 3.10 or later.
• Secrets can be used for gRPC/HTTP export headers. High-churn labels and annotations are excluded from collection by default.