Telemetry pipeline management, OpenTelemetry, and observability cost control — from the team building LinkMesh.
LinkMeshPlan a central observability platform across on-premises and Azure: ownership, Collector blueprints, change control and measurable pilot acceptance criteria.
Roman Hüsler6 min read
LinkMeshA forwarder, a Datadog Agent and a OneAgent on every host. How to consolidate monitoring agents into one OpenTelemetry Collector — and what you lose.
Roman Hüsler13 min read
LinkMeshServices drift on attribute names until your queries quietly miss data. How to normalize OpenTelemetry semantic conventions at the collector, fleet-wide.
Roman Hüsler10 min read
LinkMeshMost sizing guides get CPU right and memory wrong. A ratio-based method for sizing an OpenTelemetry gateway, and why scaling out won't fix memory.
Roman Hüsler13 min read
LinkMeshFleet management pushes configuration out to collectors. Getting telemetry in is the other half of the job — and it is where the hours actually go.
Philippe Braxmeier9 min read
LinkMeshBindplane and LinkMesh both manage OpenTelemetry collector fleets over OpAMP. Where the control plane runs is what actually separates them.
Philippe Braxmeier10 min read
LinkMeshRegulated environments buy evidence, not dashboards. What actually changes when monitoring becomes observability — and what a FINMA audit asks for.
Philippe Braxmeier6 min read
LinkMeshPod logs carry PII, namespaces are tenants, and the audit log is evidence. What changes when a cluster's telemetry has to survive a FINMA or ISO review.
Roman Hüsler6 min read
LinkMeshA Splunk-to-Sentinel move only pays off if data volume changes on the way. What a lift-and-shift costs, and where the pipeline layer belongs.
Roman Hüsler7 min read
LinkMeshvCenter alarms, Aria dashboards and VMware Tools don't come with you. What each one maps to on Proxmox, what has no equivalent, and what to build instead.
Roman Hüsler5 min read
LinkMeshData residency is not sovereignty. Who can compel access to your telemetry, who operates the control plane, and how to keep the answer yours.
Philippe Braxmeier6 min read
LinkMeshThe javaagent instruments 100+ libraries with one flag — and ships noisy defaults. Startup cost, sampling, log bridging, and what to turn off.
Roman Hüsler6 min read
LinkMeshProxmox has no native Prometheus endpoint. How to collect node, guest, Ceph and backup signals with one OpenTelemetry Collector — and keep it on-prem.
Roman Hüsler7 min read
LinkMeshElastic Agent and Fleet versus the OpenTelemetry Collector as your collection layer — integrations and ECS against portability and OpAMP.
Philippe Braxmeier6 min read
LinkMeshEvery OpenStack service can be up while a launch takes four minutes. The difference between watching services and following a request across them.
Roman Hüsler6 min read
LinkMeshA reference architecture for telemetry under revDSG and FINMA: classify at the edge, a Swiss gateway tier, split destinations, a control plane you host.
Philippe Braxmeier7 min read
LinkMeshPII and client-identifying data are different legal categories. No scanner reliably finds the second — classify before data leaves the network.
Philippe Braxmeier10 min read
LinkMeshBoth. Adopting OpenTelemetry trades vendor lock-in for fleet operations. What that burden actually consists of, and which parts are avoidable.
Philippe Braxmeier5 min read
LinkMeshGrafana Fleet Management went GA for upstream OTel Collectors in July 2026. Where your control plane runs is the question that decides between them.
Philippe Braxmeier9 min read
LinkMeshRarely a real either/or — but their telemetry could not be more different. Sources, correlation, tenancy, and one collector fleet across both.
Roman Hüsler5 min read
LinkMeshThe Collector stopped being a protocol shim and became the enforcement point for cost, privacy and routing. What that shift changes about operating it.
Philippe Braxmeier5 min readLinkMeshMonitor Proxmox VE nodes, VMs, storage and Ceph with OpenTelemetry and Grafana. Follow the setup guide with downloadable Collector config and dashboard.
Roman Hüsler15 min read
LinkMeshAttribute-based routing in the OpenTelemetry Collector — a priority-ordered, first-match-wins cascade you build with a picker instead of raw OTTL.
Roman Hüsler9 min read
LinkMeshKeep credentials out of committed Collector config: env vars, file substitution, Kubernetes Secrets, Vault — and how to rotate without downtime.
Roman Hüsler10 min read
LinkMeshEvery other data domain grew a middle tier. Observability never did — because agents shipped with backends. What that historical accident still costs.
Philippe Braxmeier6 min read
LinkMeshObservability explores telemetry; AIOps applies ML to detect, correlate and automate. Where the line sits, and why your pipeline decides both.
Philippe Braxmeier12 min read
LinkMeshReal logs aren't JSON. How to fold multiline stack traces into single events, parse just enough to query them, and ship them to Loki.
Roman Hüsler5 min read
LinkMeshInstall the OpenTelemetry Collector as a Windows service, read the counters and event channels you already know, and send them to Grafana.
Roman Hüsler9 min read
LinkMeshCollect patch state — pending updates, security patches, reboot-required — from Windows and Linux with one OTel Collector, and alert in Grafana.
Roman Hüsler7 min read
LinkMeshEvery backend cost lever acts on data you already transmitted. Where the spend is really decided, and why commitment tiers make it permanent.
Philippe Braxmeier6 min read
LinkMeshRunning one Collector is easy; running a fleet is a discipline. Drift, version skew and blind rollouts — and how to manage them from one control plane.
Philippe Braxmeier7 min read
LinkMeshVersion drift, unpatched CVEs and 3am rollbacks. How to detect outdated OpenTelemetry collectors and upgrade them health-gated and reversibly.
Roman Hüsler5 min read
LinkMeshThe 10 storage metrics worth tracking — capacity, performance, health — and how to collect them from servers and arrays with one OTel Collector.
Roman Hüsler6 min read
LinkMeshEditing a pipeline blind is how a greedy regex reaches production. Open any processor and watch a real record go in one side and out the other.
Philippe Braxmeier4 min read
LinkMeshSplunk, Sentinel, Grafana and Dynatrace in one estate is normal, not a mistake. Collect once and fan out, instead of one agent per backend.
Roman Hüsler6 min read
LinkMeshMost tools tell you your telemetry volume in next month's invoice. Put a live records-per-second number on every edge of the topology instead.
Philippe Braxmeier6 min read
LinkMeshCollector running, no data arriving, logs silent. How to debug a telemetry pipeline by inspecting real records per route, processor by processor.
Roman Hüsler7 min read
LinkMeshMask PII in-pipeline before telemetry crosses your network boundary — attribute deletion, regex masking, and hashing that preserves correlation.
Roman Hüsler6 min read
LinkMeshDual-ship telemetry to two backends to validate a migration — without double-counting hosts, doubling metrics or duplicating every log line.
Roman Hüsler8 min read
LinkMeshRoll out Collector config safely across a fleet: validate, canary to a subset, version as source of truth, and roll back fast with OpAMP.
Roman Hüsler7 min read
LinkMeshConfig drift is silent: unredacted PII, cost blowouts, unattributed data. How to detect it against your source of truth — and prevent it with OpAMP.
Roman Hüsler8 min read
LinkMeshTurn telemetry policy — PII redaction, cost caps, approved destinations — into rules enforced across a Collector fleet with OpAMP and a control plane.
Philippe Braxmeier9 min read
LinkMeshSeven Cribl alternatives compared for 2026 — self-hosted, OpenTelemetry-native and SaaS — with an honest look at when Cribl is still the right call.
Philippe Braxmeier10 min read
LinkMeshReplace Filebeat, Metricbeat and Fleet-managed Elastic Agent with OTel Collectors and OpAMP — and keep Elasticsearch and Kibana while you dual-run.
Roman Hüsler13 min read
LinkMeshRun OpenTelemetry alongside OneAgent, validate signal parity, then cut over — without losing the coverage OneAgent gave you automatically.
Roman Hüsler11 min read
LinkMeshRetire universal and heavy forwarders and HEC for OTel Collectors, keep Splunk as a destination while you dual-run, and move cost off per-GB.
Roman Hüsler12 min read
LinkMeshHow much work a New Relic exit really is: swap the APM and Infrastructure agents for OTel SDKs and Collectors, keep New Relic on OTLP, then cut the bill.
Roman Hüsler13 min read
LinkMeshWhat a Datadog exit actually takes: replace the Agent with OTel Collectors, dual-ship to compare signals, then cut the per-host and per-GB bill.
Roman Hüsler11 min read
LinkMeshTail vs head sampling in OpenTelemetry: always keep errors and slow traces, sample consistently by trace ID, and govern sample rates centrally.
Roman Hüsler9 min read
LinkMeshIsolation, tenant identification and per-tenant routing for a shared OpenTelemetry Collector fleet — dedicated vs shared pipelines, quotas, RBAC.
Roman Hüsler11 min read
LinkMeshA single Collector is a single point of failure. HA across agent and gateway tiers: load-balanced pools, sending queues, and persistent queues.
Roman Hüsler8 min read
LinkMeshTreat Collector configs like code: git-backed storage, commit and publish, review and rollback — plus the drift detection plain GitOps can't give you.
Roman Hüsler6 min read
LinkMeshA practical playbook to cut observability spend: measure per-route volume, filter noise at the edge, sample, and route before the invoice lands.
Philippe Braxmeier7 min read
LinkMeshOpenTelemetry replaces collection and transport, not your backend. What it takes over cleanly, what you must rebuild, and how to scope a migration.
Philippe Braxmeier7 min read
LinkMeshOTLP, the Collector and OpAMP make collection portable and give you a real exit path. What OpenTelemetry decouples — and what it doesn't.
Philippe Braxmeier11 min read
LinkMeshOpAMP remote-manages a fleet of OpenTelemetry collectors — status up, config down. How the protocol works, and the control-plane gap it leaves.
Roman Hüsler7 min read
LinkMeshAlloy is an OpenTelemetry Collector distribution. The real differences: config language, remote management, ecosystem — plus how to run both in one fleet.
Philippe Braxmeier7 min read
LinkMeshFive ways to deploy an OpenTelemetry Collector — agent-per-host, DaemonSet, sidecar, gateway, hybrid — compared, plus the no-collector option.
Roman Hüsler11 min read
LinkMeshThe reference architecture for OpenTelemetry at scale: agent collectors, a central gateway tier, and OpAMP as the control plane that configures it.
Philippe Braxmeier11 min read
LinkMeshA telemetry pipeline sits between your systems and your backends — where you control cost, routing and PII. What it is and how to run your own.
Philippe Braxmeier10 min read
LinkMeshA control plane for your whole OpenTelemetry Collector fleet: enroll every node over one port, push config from one place, watch throughput live.
Philippe Braxmeier3 min read