Skip to main content
TL;DRThis guide shows you how to send telemetry to ClickHouse Observability from OpenTelemetry Collectors managed by Telflo. You’ll learn how to:
  • Build a collector configuration in Telflo’s visual editor or using the CLI
  • Send telemetry to fully managed or self-managed ClickHouse Observability
  • Remotely deploy configurations to a fleet of collectors
  • Confirm the telemetry is available in ClickHouse Observability
Time required: 15-20 minutes

​
What is Telflo?

Telflo is a vendor-neutral control plane for OpenTelemetry Collectors, providing centralized management for collector configuration across a fleet. It makes large fleets manageable with a visual editor, pipeline tests, and remote deployment of configurations. Telflo manages standard upstream OpenTelemetry collector builds, with no Telflo-specific distribution and no proprietary agent, so collectors your teams already run stay on the same build. Telflo is also accessible outside the browser, through a command-line interface. Pair that with the ClickStack MCP server and the same agent that queries your telemetry can fix the pipeline that produced it.

​
Why Telflo + ClickHouse Observability?

Fully managed ClickHouse Observability handles ingestion, schemas, storage, and scaling. With self-managed ClickHouse Observability, you control ingestion and database configuration while ClickHouse Cloud manages the underlying infrastructure. Telflo manages the OpenTelemetry Collectors that send telemetry from your hosts, clusters, and regions to either deployment. Telflo provides:
  • Centralized management of OpenTelemetry Collectors across hosts, containers, and Kubernetes
  • Visual pipeline editing, with controlled rollouts to the whole fleet or a single group
  • Changes verified against sample telemetry before they reach production collectors
  • Routing the same telemetry to ClickHouse Observability and other destinations at the same time
  • Fleet visibility: which collectors are online, what each one is running, and which have drifted
Key takeaways
  • ClickHouse Observability provides the storage and investigation experience
  • Telflo operates the collector fleet that feeds it

​
Prerequisites

​
Integrate ClickHouse Observability with Telflo

1

Build a collector configuration

  1. Sign in to Telflo. On the Configurations list, click New Configuration, then Blank configuration.
  2. In Name Your Configuration, enter a name and click Create.
  3. Add receivers: drag one from the palette onto the canvas for each kind of telemetry you collect.
  4. Add processors: Memory Limiter and Batch, plus any others you need to shape data before it leaves the collector. See the full list of processors Telflo supports.
2

Add the exporter

Choose the exporter for your deployment. Fully managed ClickHouse Observability accepts telemetry through its managed OTLP endpoint. For self-managed deployments, you can write directly to ClickHouse or send OTLP to a collector you already run.
Use the managed OTLP endpoint. ClickHouse handles ingestion, schemas, storage, and scaling; the collectors managed by Telflo collect and forward telemetry from your environment.
  1. Drag the OTLP/HTTP exporter onto the canvas.
  2. Set Endpoint to your managed OTLP HTTP endpoint, including port 4318, for example https://<service-id>.otel.<region>.aws.clickhouse.com:4318.
  3. Under HTTP Headers, click Add header, then enter Authorization as the Header name and Bearer ${env:CLICKSTACK_INGESTION_KEY} as the Header value. Telflo asks for the environment variable value when you install the collectors. Supply your ingestion API key as CLICKSTACK_INGESTION_KEY, without the Bearer prefix.
Use the OTLP exporter, not the direct ClickHouse exporter, for fully managed deployments.
Draw edges from the receivers through the processors to the exporter, one pipeline per signal. The status row reports config valid or the number of errors to fix, and the YAML tab shows the generated configuration.Before you roll anything out, you can run sample telemetry through the configuration and check what comes out the other end. See testing in Telflo. Alternatively, you can describe what you’re trying to do and the Telflo AI assistant, opened with the AI button in the toolbar, builds, validates, and tests the configuration for you. Your own agent can do the same through the Telflo CLI.
3

Publish a version

Click Publish in the toolbar. Published versions are read-only; to keep iterating, click Create New Version.
4

Create a fleet and roll out

  1. Go to Fleet, click New Fleet, and choose where your collectors run, on VMs or in Kubernetes. Enter a Name and click Create fleet.
  2. On the fleet’s Deploy Groups tab, click Assign on the group you want to deploy to and choose your published configuration and version. Start with a small group to verify ingestion before rolling out to the rest of the fleet.
  3. Click Install on the group and enter a value for each environment variable your configuration references; Telflo adds them to the install command. Run that command where the collectors should run, following Telflo’s installation instructions. It carries the fleet’s install token along with your value, so treat it as sensitive.
  4. On the Collectors tab, collectors report Pending pickup, then Applying, then In sync.
To roll out a change, publish a new version and assign it to the group; to roll back, assign a previously deployed version.
5

Verify telemetry

Open ClickHouse Observability for your service in the ClickHouse Cloud console. Search for one of your services in the Search view; use Chart Explorer for metrics. Check that the expected logs, traces, and metrics have arrived before rolling the configuration out to the rest of your fleet.

​
Advanced configuration

​
Fan-out routing

The same telemetry can go to ClickHouse Observability for long-term storage, to another platform during a migration, and to a SIEM for security analysis at once. Add the exporters on the canvas and connect them to the pipelines that should feed them, or start from a prebuilt pipeline in Telflo’s flow library.

​
Compression and performance

  • OTLP/HTTP and OTLP exporters: keep Compression at gzip.
  • ClickHouse exporter (self-managed only): keep Async Insert on, leave Compression at lz4, and add a Batch processor upstream. See Optimizing inserts.

​
Next steps

  • Build dashboards: create visualizations in ClickHouse Observability for logs, metrics, and traces
  • Set up alerts: configure alerts in ClickHouse Observability for critical conditions
  • Refine with the Telflo AI assistant: ask it to adjust a pipeline or propose tests
  • Coming from another vendor: flows cover the cutover, replacing the agents already on your fleet and dual-shipping to both backends while you compare them

​
Read more