DataBuff v0.1.9 is ready: SkyWalking 8.0–8.3 agents can hit the existing ingest port; LLM providers have full CRUD; agent shell and read-only SQL cannot reach credential tables. Versus v0.1.8: 10 commits, no schema migration.
Highlights
- SkyWalking 8.0–8.3: unqualified gRPC service names for Trace / JVM / Management are accepted on the existing ingest port.
- LLM provider CRUD: create, update, delete, and detail views in config management.
- Credential access hardening: AI agent shell and read-only Doris queries cannot reach
config_llm_provider/api_key_cipher. - AI brain: duplicated streaming messages are fixed.
- Doris-down LLM key: an in-memory LLM provider still takes effect when the config store is unavailable.
- Service inspect:
inspectServiceusesLIMIT 1existence probes and reuses HTTPcomponentTypes. - Docs: Python OTLP integration quick-start (GitHub #75).
Install & upgrade
Fresh install: Docker install. Existing installs: upgrade & uninstall. No schema migration this release (schema_version stays 8). Kubernetes does not support in-place upgrade; uninstall then reinstall.
# Fresh install (wipes install dir and data/)
curl -fsSL https://databuff.ai/databuff/ai-apm-install.sh | bash
# In-place upgrade (keeps observational data/)
curl -fsSL https://databuff.ai/databuff/ai-apm-update.sh | bash