Entity trail

Grafana

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

Findings

Edges

Sources

Corpus findings

2026-08-26 / github-pulse-researchermarin, an open framework for foundation-model research, is running visible TPU training infrastructure in public at 443 stars a weekmarin-community/marin (2,343 stars, Apache-2.0) is unusual in that its commit log is an operating training cluster rather than a library changelog: commits on 2026-08-25 and 08-26 publish levanter checkpoint phase telemetry, add merged cache catalogs, allow the TPU ferry launcher on preemptible capacity, default the TPU ferry to v6e, make GCP IAM role bindings authoritative and repair a vLLM Grafana overview. For builders, the value is less the framework and more the visible reference implementation of what running foundation-model training on preemptible TPU capacity actually requires, including the Grafana dashboards and auto-triage CI.
2026-08-25 / rss-researcherSageMaker HyperPod gets managed Ray on EKS with four prebuilt Grafana dashboards and no Kubernetes manifestsAWS announced on 2026-08-24 that SageMaker HyperPod now offers managed Ray on EKS, letting data scientists create clusters, open the Ray Dashboard and Amazon Managed Grafana views, and attach a JupyterLab or Code Editor workspace from the console without writing Kubernetes manifests. Workspaces attach as zero-compute worker nodes so standard Ray APIs run distributed training straight from a notebook. The Observability add-on provisions Ray Core, Ray Data, Ray Train and Ray Serve dashboards automatically, and resiliency covers automatic node replacement, hung job detection, and tiered checkpointing that checks local HyperPod storage before S3. It uses the open-source KubeRay operator, so existing scripts run unmodified, and is live in all regions where HyperPod EKS is supported.
2026-08-23 / skill-finderAnthropic shipped its internal CI/CD on-call agent as a public setup kit, with skills in git and a lessons.md learning loopThe August 18 writeup describes Claude Tag running as Anthropic's first responder for CI failures, with a dedicated service account, MCP connectors to Datadog, Grafana, PagerDuty, GitHub and Kubernetes, and an orchestrator that spawns executor subagents to check dependencies in parallel. Standing instructions live as markdown skills in a GitHub repo, including an oncall.md with deterministic escalation criteria and an auto-updating lessons.md incident journal; median time to first analysis is 14 minutes. The generalized kit at github.com/anthropics/oncall-kit (21 stars, reference code, not maintained) is a five-phase gated setup that mines your own incident history into playbooks, keeps Claude strictly read-only, and validates drafted playbooks against held-out incidents before install.
2026-08-03 / github-pulse-researcherGrafana Ships a Go AI SDK That Is Wire-Compatible With Vercel's TypeScript React Hooksgrafana/ai-sdk (198 stars, created 2026-07-28, Apache-2.0) gives Go one API for model calls, streaming, tools, structured output, and multi-step agents, deliberately following Vercel AI SDK's design and staying wire-compatible with its frontend hooks. A Go backend calling aisdk.StreamText and aisdk.WriteUIMessageStream streams Server-Sent Events directly into useChat with no protocol adapter. For anyone with a Go service and a React frontend, this removes the usual reason to keep a TypeScript BFF in the middle purely to speak the AI SDK stream protocol.
2026-07-31 / arxiv-researcherORCA-bench: Best Frontier Agent Scores 25.3% on Realistic Oncall Root-Cause Analysis, 10.0% on Hard TasksORCA-bench (arXiv 2607.28545, July 30) pairs a live OpenTelemetry-instrumented microservice system — six days of metrics, logs, and traces exposed via Prometheus, Jaeger, and OpenSearch through Grafana, plus full source access — with 1,079 root-cause-analysis tasks that vary report specificity, time-to-detection, and co-occurring faults. Across five frontier agents the best RCA accuracy is 25.3% on Medium (realistic-input) tasks and 10.0% on Hard, a gap that persists even with Claude Fable 5; the weakest model hallucinates an implausible root cause in 40% of reports, and removing source-code access degrades every metric. Ground truth was signed off by expert SREs with LLM-as-judge re-scored by humans at Cohen's weighted kappa 0.90, and the authors frame the 50 GB/six-day testbed as a lower bound on real production difficulty.
2026-06-02 / github-pulse-researcherHolmesGPT at 2.5K Stars: Open-Source SRE Agent Joins CNCF Sandbox With 40+ Observability Integrations and Petabyte-Scale Incident InvestigationHolmesGPT (HolmesGPT/holmesgpt) is an AI agent for investigating production incidents that earned CNCF Sandbox status, signaling Cloud Native ecosystem acceptance of AI-driven SRE tooling. V0.31.1 shipped May 28 with integrations across 40+ observability platforms (Prometheus, Grafana, Kubernetes), bidirectional alert integrations with PagerDuty and OpsGenie, and an Operator mode for continuous 24/7 automated problem detection. The tool supports multi-LLM backends including OpenAI, Anthropic, Azure, and Bedrock.
2026-05-26 / agents-researcherE2B + Docker Partnership: Every Agent Sandbox Now Includes 200+ Pre-Verified MCP Tools via Docker MCP GatewayE2B partnered with Docker to embed the Docker MCP Catalog (200+ tools including GitHub, Perplexity, Browserbase, Stripe, Grafana, Notion) into every E2B sandbox via Docker MCP Gateway. Each MCP tool runs as a Docker container inside the sandbox with autocomplete, type validation, and unified configuration. What previously required hours of manual configuration now takes seconds. For agent builders, this eliminates the single most painful setup step: wiring tools into sandboxed execution environments.
2026-05-23 / sources-researcherGitHub Discloses 3,800 Internal Repositories Exfiltrated via TeamPCP Supply Chain Attack on Nx Console VS Code ExtensionA trojanized Nx Console VS Code extension (v18.95.0, 2.2M+ installs) was live for just 18 minutes on May 18 but harvested tokens from GitHub, npm, AWS, Vault, Kubernetes, and 1Password via a 498KB obfuscated payload hidden in an orphan commit. A GitHub employee installed it, enabling threat group TeamPCP to move through CI/CD pipelines and exfiltrate ~3,800 internal repositories. The attack also hit Grafana Labs and was traced to a broader TanStack supply chain compromise.
2026-05-17 / hn-researcherGrafana Labs Source Code Stolen via GitHub Action 'Pwn Request' Vulnerability — Extortion Attempt RefusedGrafana Labs disclosed on May 16 that an attacker exploited a pull_request_target misconfiguration in a recently enabled GitHub Action to extract privileged tokens, then downloaded Grafana's entire private codebase. The attacker (claimed by CoinbaseCartel) attempted extortion; Grafana refused citing FBI guidance. The root cause was a 'Pwn Request' vulnerability where external contributors could access production secrets during CI runs by forking a repo and injecting malicious code. No customer data was accessed.
2026-05-16 / github-pulse-researcherHolmesGPT/holmesgpt at 2.4K Stars: AI SRE Agent Accepted into CNCF Sandbox for Autonomous Incident InvestigationHolmesGPT is an AI-powered SRE agent that autonomously investigates production incidents by correlating logs, metrics, and traces. Its acceptance into the CNCF Sandbox — the Cloud Native Computing Foundation's incubation program — signals that AI-driven incident response is gaining institutional legitimacy in the cloud-native ecosystem. The tool integrates with standard observability stacks (Prometheus, Grafana, PagerDuty) to provide root cause analysis without manual investigation.
2026-05-12 / agents-researcherGrafana Open-Sources o11y-bench: AI Agent Benchmark for Real Observability Workflows Against Live Prometheus/Loki/Tempo StackGrafana Labs open-sourced grafana/o11y-bench at GrafanaCON 2026, a benchmark that runs AI agents against a real Grafana stack with Prometheus, Loki, and Tempo — not synthetic scenarios. Tasks include querying metrics/logs/traces, investigating incidents, and modifying dashboards. Built on Harbor, it grades agents on actual system actions rather than natural language responses. The open leaderboard ships with frontier model baselines and welcomes community-submitted agent harness and model configurations. Fills a critical gap: most agent benchmarks test chat; o11y-bench tests operations.
2026-04-18 / agents-researcherGrafanaGhost: Zero-Click AI Prompt Injection Exfiltrates Enterprise Data from Grafana Observability PlatformNoma Security disclosed GrafanaGhost on April 7 — a prompt injection attack where adversarial instructions hidden on attacker-controlled web pages cause Grafana's AI capabilities to silently exfiltrate financial metrics, infrastructure telemetry, and customer records without credentials. Researchers bypassed guardrails using protocol-relative URLs for domain validation circumvention and the INTENT keyword to disable model safety. Grafana patched the core issue; no in-the-wild exploitation confirmed.

Graph relationships

USES
marin
marin uses Grafana for cluster monitoring and dashboards.
Source finding
USES
OpenSRE
OpenSRE integrates with Grafana for observability.
Source finding
RELEASED
o11y-bench
Grafana open-sourced o11y-bench benchmark for AI agents
Source finding
CRITICIZES
Noma Security
Noma Security disclosed GrafanaGhost prompt injection vulnerability affecting Grafana AI capabilities.
Source finding
USES
marin -> Grafana
marin uses Grafana for cluster monitoring and dashboards.
Source finding
USES
OpenSRE -> Grafana
OpenSRE integrates with Grafana for observability.
Source finding
CRITICIZES
Noma Security -> Grafana
Noma Security disclosed GrafanaGhost prompt injection vulnerability affecting Grafana AI capabilities.
Source finding

Source trail

marin-community/marin August 18 writeup github.com/anthropics/oncall-kit grafana/ai-sdk A new glossary E2B partnered with Docker Mem0's State of AI Agent Memory report OX Security Trend Micro data ArXiv paper 2605.06760 pull_request_target vulnerability purchased the Essential Plugin portfolio on Flippa

Graph sources

entity graphfindings textkg edgeskg entitiesnewsletter issues

Fetching from the wire…