Databuff
GitHub
Skip to content

中文  |  English

Roadmap

Delivered · v0.1

Open-source AI-native OpenTelemetry APM — the core loop is working end to end.


Next Phase · Core Plans

Five priorities to complete the full collect → observe → intelligently act pipeline:

#DirectionGoal
One-Agent open sourceRelease a unified collection agent for traces, metrics, logs, and host metrics — ready to connect to DataBuff Ingest out of the box
OpenTelemetry logsIngest OTLP Logs, support OpenTelemetry log collection and trace correlation, completing the three pillars of observability
eBPF zero-instrumentation APMKernel-level collection via eBPF with zero code changes for host, container, and K8s scenarios — covering runtime areas agents cannot reach
Agent observabilityObserve AI agents themselves — call chains, token usage, tool calls, latency, and errors — so agent runs are traceable and diagnosable
AI agent memoryPersistent memory for in-platform multi-agents — retain context and diagnostic conclusions across sessions so AI learns your system over time

Near-term Plans

DirectionGoal
Deeper AIStreaming chat, more digital experts, causal root-cause analysis
Richer APMLog correlation, RUM, broader middleware coverage
Stronger alertingComposite rules, incident workflows, notification integrations
More stable deploymentMulti-node clusters, persistence options, Helm Chart

Long-term Vision

From "AI-assisted viewing" to "AI-autonomous management" — that is DataBuff's ultimate direction.