Running AI features reliably
AI SLOs and alarms
AI feature SLOs and alarm rules
- Observability: Logs, Metrics and Traces
- From Alerts to Runbooks: Wiring a Pager That Points to a Fix
- Alerting on Agent Failures Without Paging for Every Error
- Dashboards for AI Features: The First Screen
- Latency Budgets for an AI Feature: Where the Milliseconds Go
- Queue Depth, Not Just Connections: Metrics for a Model Gateway
- SLOs for a Model-Backed Endpoint: Choosing the Right SLI
Tracing agent runs
Tracing agent and model runs
- Span Attributes That Let You Slice by Model, Tenant and Outcome
- Structuring Logs and Traces for AI Agents: What to Record
- Tracing the Query Path: Spans for Retrieval, Reranking, Calls
- A Trace Attribute for Remaining Time at Each Tool Call
- Building a Trace Viewer That Answers the First Question
- Building the Agent Status Page Your Support Team Reads
- Choosing a Tracing Backend for Agent Runs
- Finding the Failing Step in a Trace Without Reading Logs
- All of Tracing agent runs
Logs and audit trails
Logs, events and audit trails
- Logs, Events and Spans: Picking the Right Signal per Step
- A Rule Change Log That Reconstructs a Past Call
- Explaining an Agent-Issued Refund to the Customer
- Log Retention for Agent Traces: What to Keep and for How Long
- Paging the Audit Log: Immutable Rows, Replays and Retention
- Redacting Prompts and Outputs Before You Log Them
- The Approval Audit Row: Fields to Keep After the Screen Closes
- What the Audit Log Must Contain to Prove a Query Was Read-Only
- All of Logs and audit trails
Incidents and runbooks
Incidents, runbooks and postmortems
- When an Agent Fails at 2 A.M.: A Runbook for Agent Incidents
- After the Postmortem: Tracking Action Items Until They Ship
- Paging on a Failed Money Probe: Runbooks and Ownership
- Postmortems That Change Code, Not Just Documents
- Reproducing an Agent Incident from Traces in Staging
- When a Model Stops Mid-Sentence: A Truncation Repair Runbook
- Who Gets Woken When a Write Path Fails: Runbooks and Ownership
On-call for AI systems
On-call for AI systems
Cost and capacity
Cost, latency and capacity operations
- Comparing Reserved Capacity Across Two Vendors
- Keeping Model Calls Warm: Reserved Capacity
- Negotiating a Reserved Capacity Contract: Terms That Bite
- A Commitment Review: What to Bring and When to Renegotiate
- A Customer-Facing Statement for Agent Spend: Line Items
- A Spill Plan: What to Shed When Reserved Capacity Fills
- Forecasting Model Spend from Traffic Percentiles
- Keeping Two Model Versions Warm Without Doubling Spend
- All of Cost and capacity