Standardise the fleet
Agree resource attributes, source settings and destination rules. Use collector groups to apply shared configuration to the intended environments.
Self-hosted OpenTelemetry control plane
Manage OpenTelemetry centrally. In your infrastructure.
Standardise collector configurations and telemetry pipelines across on-premises, Azure and Kubernetes. Give platform teams a shared management layer while retaining control over deployment and telemetry destinations.
25 Collectors free after a no-card registration · 5 without one How the free tier works
A shared operating model for a hybrid fleet
For platform and observability teams coordinating application, security and operations requirements. Agree ownership and repeatable configuration before scaling a central observability platform across environments.
Make changes traceable
Inspect configuration history and differences. Define who can save changes and how your team verifies the result. Saving applies configuration; it is not a draft awaiting approval.
Onboard one workload first
Start with one on-premises workload and one Kubernetes workload. Verify the supported collector runtime, connectivity and data arriving at your backend before extending the blueprint.
Separate management traffic from workload telemetry
Allow each collector to reach the control-plane endpoint and its configured data destinations. Design TLS, proxies, identity and failure boundaries explicitly. Diagnostic sample capture is a separate workflow that also needs a data-handling review.
Inspect the fleet before changing it
What belongs in your architecture review
Hosting and data paths
You operate the control plane. Collectors send workload telemetry to the destinations you configure; cloud destinations still involve cloud egress. Inventory, configuration, health and collector self-telemetry belong in the management-plane review.
IAM and automation
Validate roles, group scopes, OIDC and service-account requirements against the current edition and documentation. Do not equate central management with an approved enterprise IAM integration.
GitOps and rollout
The external Git repository is a history mirror, not an inbound deployment source. Use separate test groups and an explicit change process. Do not assume automated progressive rollout or pull-request approval gates.
Rollback scope
Validate the licensed rollback operation. Version history does not restore source/destination settings or secret values. Record those separately and rehearse recovery before production changes.
ITSM, CMDB and dashboards
Keep incident workflows, service ownership and dashboard provisioning in the architecture as explicit integrations. LinkMesh manages collectors and pipelines; it does not replace the observability backend or ITSM platform.
Offline operation and key management
An offline design needs package mirrors, certificate and licence processes, backups and reachable destinations. BYOK/KMS integration and regulatory acceptance require specific evidence; self-hosting alone proves neither.
Make the pilot answer a buying decision
- Inventory runtimes, owners, signals and network boundaries; agree success thresholds before starting.
- Connect an on-premises and a Kubernetes test collector; measure setup effort and document each manual step.
- Check required attributes, event counts and sensitive synthetic test values at the destination.
- Apply one change, verify its effective result and rehearse the full rollback, including settings outside version history.
- Interrupt the control connection and a destination independently; measure continued collection, buffering, loss and recovery.
- Review the evidence with platform, application, security and operations owners before expanding.
Common questions
Can one control plane manage on-premises and Azure collectors?
Evaluate that design with supported collector versions and approved network routes. AKS and OpenShift workload collection also depends on cluster permissions, security policies and the selected receivers; a generic Kubernetes diagram is not a compatibility certificate.
Does self-hosting guarantee compliance?
No. It gives you deployment control. Your organisation must assess data classification, access, retention, encryption, operations and external dependencies against its requirements.
Does LinkMesh replace an observability platform?
It provides the collector-management layer. Storage, queries, dashboards, alerting and service-management workflows remain responsibilities of the selected backends and connected systems.
Evaluate with your architecture and your workloads
Bring your collector inventory, target backends, network boundaries and required controls. Pricing is based on managed collectors; verify the edition needed for automation, SSO and rollback.
Start with one test collector.
Save the current configuration, connect one collector and verify a real signal in your backend. Test one change and its rollback before expanding to the fleet.
Setup guides and evaluation checklist
- Install and verify your first data flow →
- Set up HTTPS and the reverse proxy →
- Download the evaluation checklist
- Check versions and limitations
- Ask a technical question
No email required for the download. Includes failure tests, rollback and checks for ingest-cost assumptions.