Observability

48 posts tagged observability.

Nov 10, 2026

How to Run Kafka on Kubernetes with Strimzi 1.3: KRaft Setup (2026)

Step-by-step Strimzi 1.3.1 guide: install the operator with Helm, run a KRaft Kafka 4.3.2 cluster with KafkaNodePools, manage topics and users as CRDs, connect over mTLS, add Prometheus alerts, upgrade safely and fix under-replicated partitions and consumer lag.

13 min
Nov 10, 2026

How to Self-Host SigNoz with OpenTelemetry: Docker or Helm (2026)

Step-by-step SigNoz v0.149.0 self-hosting with Foundry on Docker or the 0.149.0 Helm chart, an OpenTelemetry Collector 0.165.0 agent config, Python auto-instrumentation, log-trace correlation, retention, alerts, and a no-data troubleshooting table.

12 min
Nov 10, 2026

How to Triage Logs With Drain3 and a Local LLM (Ollama, 2026)

Build a log triage pipeline in Python: mine templates with Drain3 0.9.11, flag new and rare templates against a baseline, redact, and send a 629-token summary (not 25,509 tokens of raw logs) to a local Ollama model. Tested on Loghub HDFS and Linux samples.

15 min
Nov 10, 2026

How to Manage Grafana Dashboards as Code: Terraform, Git Sync (2026)

Step-by-step Grafana dashboards as code for Grafana 13: generate JSON with the Foundation SDK or Grafonnet, ship it with provisioning files, the Terraform provider 4.54 or Git Sync, lint it in CI, set folder permissions, and fix common provisioning errors.

12 min
Nov 10, 2026

Golden Signals vs RED vs USE Metrics Explained (2026): PromQL Examples

The four golden signals, RED and USE explained for 2026: which framework fits services and which fits resources, PromQL you can paste, histograms vs summaries, native histograms stable since Prometheus 3.8, SLIs and on-call dashboards.

12 min
Nov 10, 2026

ClickHouse for Observability in 2026: ClickStack, SigNoz, 431 PiB of Logs

Why ClickHouse won logs and traces: columnar compression of 16x to 17x, ClickStack after the HyperDX deal, SigNoz v0.149 and Uptrace compared, and real cost math against Datadog's $1.70 per million indexed events and Elasticsearch logsdb.

11 min
Nov 10, 2026

AI Log Analysis in 2026: What Works, What Hallucinates, What It Costs

AI log analysis in 2026, tested: Drain-style templating cut 500 Loghub lines to 629 tokens, the best LLM agent on OpenRCA found 11.34% of root causes, and raw logs cost about $723 per GiB at Sonnet 5.5 prices. What Datadog, Elastic and Grafana ship.

11 min
Oct 31, 2026

How to Trace an LLM App with OpenTelemetry and Langfuse (2026)

Self-host Langfuse v4 with Docker Compose, instrument a Python tool-calling agent with the OpenTelemetry-based Langfuse SDK v5, track token cost, sessions and user feedback scores, route other services through an OTel Collector, and fix traces that never show up.

13 min
Oct 31, 2026

Structured Logging Best Practices in 2026: JSON, Trace IDs, PII, Cost

Structured logging best practices for 2026: which JSON fields to emit, how log levels map to OpenTelemetry's 24 severity numbers, putting trace_id and span_id in every line, what never to log, sampling without losing errors, real per-GB log prices, and logging for LLM apps.

13 min
Oct 31, 2026

Service Mesh in 2026: Istio Ambient vs Linkerd vs Cilium, Do You Need One?

Do you need a service mesh in 2026? Istio ambient is GA and at 1.31.3, Linkerd 2.20 stable builds come only from Buoyant, Cilium deprecated its own mutual auth for ztunnel, and backend developer mesh use fell from 18% to 8%. Overhead data, a comparison table and a decision guide.

13 min
Oct 31, 2026

How to Build Anomaly Detection Alerts in Prometheus 3 (2026)

Step-by-step PromQL anomaly detection for Prometheus 3.15: z-score recording rules with avg_over_time and stddev_over_time, week-over-week seasonal predictions, double_exponential_smoothing behind its feature flag, Grafana's promql-anomaly-detection rules, promtool tests and fixes.

13 min
Oct 31, 2026

Model Drift Explained (2026): Data vs Concept Drift, PSI, KS and LLMs

Model drift explained for 2026: data drift vs concept drift vs prediction drift, how PSI, KS and Wasserstein behave at scale, why LLM quality drifts when providers retire snapshots, which tools still exist after WhyLabs and SageMaker Model Monitor, and when to retrain.

12 min
Oct 31, 2026

How to Measure DORA Metrics From GitHub: gh api and SQL (2026)

Step-by-step: measure all five DORA metrics from GitHub with environments, the Deployments API, gh api, PR merge data and SQLite, plus revert and incident labels for change fail rate, Apache DevLake dashboards, the archived Four Keys project and a troubleshooting table.

13 min
Oct 31, 2026

LLM Observability Tools Compared (2026): Langfuse vs LangSmith vs Phoenix

Langfuse, LangSmith, Arize Phoenix, Helicone, Braintrust, Datadog and OpenLLMetry compared in October 2026: what each one bills on, list prices for a 300,000-run agent workload, licences and self-hosting, and how far the OpenTelemetry GenAI conventions have really got.

14 min
Oct 31, 2026

How to Set Up kube-prometheus-stack: Prometheus and Grafana (2026)

Install Prometheus, Alertmanager and Grafana on Kubernetes with kube-prometheus-stack 92.16.1: pinned Helm values, ServiceMonitor and PodMonitor examples, retention and disk sizing, Slack routing with a dead man's switch, dashboards as ConfigMaps, remote_write, and fixes for targets that show down.

17 min