Token + Latency Analytics: Building a Dashboard That Engineers Actually Use
Turning TTFT/TTR and cost into trends, budgets, and actionable insights across your conversations.
Deep dives on Skyflo’s architecture, safety model, and releases — written for DevOps and SRE teams shipping real systems.
Turning TTFT/TTR and cost into trends, budgets, and actionable insights across your conversations.
A single-tenant Slack bridge plan: streamed updates, approvals in-thread, and guardrails that don’t feel heavy.
A design for summarizing older turns when you approach context limits—without losing the details operators care about.
Loops, batching, parallelism, and summarization—where code beats prompts, and how to sandbox it safely.
A roadmap post: defer tool schemas until needed, reduce token bloat, and keep the agent accurate under pressure.
Adding read-only metrics tools so an agent can answer the question everyone asks first: “what’s hot right now?”
Rendering manifests with inline values, catching surprises early, and building a diff-first culture.
Shipping safe rollback primitives for deployments/daemonsets/statefulsets—and where approvals belong.
Focus, live regions, contrast, and keyboard navigation—what we changed to make a command-center UI work for everyone.
Operators don’t trust black boxes. Here’s how we expose LLM latency and usage without spamming the UI.
Why we moved, what broke, and how Streamable HTTP made MCP communication simpler and more reliable.
Designing bulk approval controls that respect read-only tools, remain idempotent, and keep the operator in control.
Why fast history, debounced search, and prompt queueing matter when you’re triaging an incident at 2am.
How Skyflo avoids leaking secrets into prompts, keeps credentials server-side, and still feels seamless in the UI.
A deep dive into the Jenkins toolset, integration-aware discovery, and why builds must be parameter-first.
What we learned rebuilding Skyflo’s core loop, and why “simpler” was the biggest performance unlock.
A hands-on guide to reliable server-sent events for long-running infra tasks, including NGINX hardening.
Why Skyflo’s MCP server exists, how tools are validated, and what “readOnlyHint” really buys you in prod.
How Skyflo compiles a compact graph, streams progress, and decides when to continue, stop, or request approval.
A practical look at approvals, safety gates, and why “agent autonomy” should still ship with guardrails.