Find articles, news, tools and templates by subject or by the work you need to do.
135 results for “Trace a slow request” · 25–36
Guide · SLOs & reliability ·
Worked request and time-based budgets accompany an SLO record, release-policy template and exception log, keeping each calculation in its original units.
Read guide: Error-budget template: calculations and release-policy fieldsNews · AI in production ·
Claude Haiku 5.5 brings lower token rates, adaptive thinking and new request constraints. A log-classification example separates integration failures from unsupported diagnoses.
Read news: Haiku 5.5 lowers prices and changes what a log classifier must handleNews · AI in production ·
Google's new image model rejects familiar generation settings. Follow a thumbnail job through its outgoing request, retries and delivery to evaluate the upgrade.
Read news: What Nano Banana 2.1 changes in an image-generation serviceNews · AI in production ·
Kolibri adds a customer-controlled bilingual model option. Its 3.46B active parameters come with about 78 GB of FP8 weights, plus memory for request state.
Read news: Aleph Alpha releases Kolibri for German and English workflowsGuide · AI in production ·
CodeRabbit brings AI feedback into pull requests, editors and the CLI. A retry loop that exceeds its caller’s deadline shows the kind of finding that can repay review time.
Read guide: CodeRabbit: how to use AI code reviewNews · Platform engineering ·
NVIDIA reports generation falling from 156.6 to 34.2 seconds on its example workload. The eight-device layout also changes how much capacity remains for other requests.
Read news: TensorRT speeds up one video by using eight GPUsGuide · AI in production ·
A timed-out scaling request needs a durable record and destination-state checks. Reusable operational skills need to recover unfinished work as well as start it.
Read guide: Recovering an interrupted agent skillGuide · Observability ·
A complete Python example combines request and error counts, while a failed regional collection explains why correct arithmetic can still produce an incomplete report.
Read guide: Collecting and aggregating operational data with PythonExplainer · SLOs & reliability ·
Worked request and time examples explain error-budget units, remaining allowance and burn rate, plus the release policy that gives them a practical role.
Read explainer: Error-budget calculations, balances and burn ratesNews · AI in production ·
Sparse routing reduces per-token compute, but the full 24B checkpoint, agent harness, license and representative workload still decide deployment fit.
Read news: What LFM2.5's 2.45B active parameters mean for local agentsExplainer · AI in production ·
Helion places CPU agent loops, token routing, GPU inference and supporting services in separate capacity zones that still share one task outcome.
Read explainer: AMD Helion separates agent execution from model inferenceGuide · AI in production ·
Connect logical tasks to every model and tool attempt, then measure cost and latency per accepted outcome without hiding failed or unknown work.
Read guide: Measuring retry amplification in AI agent workflows