Overview, Pipelines, and Collectors

Your day-to-day work in Praxis usually falls into three places: seeing whether data is flowing (Overview), defining how it should flow (Pipelines), and caring for the agents that run in your estate (Collectors). The table is a quick map; the sections below spell out what you can accomplish in each.

Area	Typical URL path	What you use it for
Overview	/overview	Answer “is my telemetry moving?”—tenant or pipeline scope, throughput views, or a welcome path to your first pipeline.
Pipelines	/pipeline	Author graphs, validate, publish versions, inspect throughput and history, and see which collectors run a given pipeline.
Collectors	/collectors	See every installed agent, health and version, bulk lifecycle actions, and per-collector metrics and pipeline context.

For shared vocabulary (management plane, collector, pipeline), see Getting started.

Overview

Use Overview when you want a fast read on operations—not when you need to edit a graph (use Pipelines for that).

• No pipelines yet — You see a welcome panel with a short value proposition, optional Create pipeline (when your role allows), and a link to external product documentation. Use this to get started with your first graph.
• At least one pipeline — You get an operational dashboard for the time range and scope you choose:
  • Scope — Choose All pipelines or a single pipeline from the filter. The URL can carry a pipeline query parameter so views are bookmarkable.
  • Time range — Presets (for example last 15 minutes through last 30 days) drive the metrics window for charts and rollups.
  • Refresh — Manual refresh clears cached overview queries; auto-refresh can re-query on an interval.
  • Summary — High-level pipeline and collector stats (for example how many collectors are reporting) for the current selection.
  • Sankey / flow — A source → … → destination style view of edge throughput so you can see how data moves through the graph. You can open insights for a source or destination node (details vary by node type) when the chart supports it.
  • Additional charts — Complementary views such as Data In & Out, Events In & Out, and a Data Distribution heatmap, all backed by the platform metrics API (Prometheus-style queries), scoped to the chosen pipeline or all pipelines.

Overview answers “what is running and how much is moving” for your tenant or a chosen pipeline. It does not replace long-term storage; telemetry still lands in the destinations you configure.

Pipelines

Pipelines is where you define how data moves: sources → processors → destinations, with validation and publish workflows that roll configuration to your collectors.

• List — Search and browse pipelines, with a time range used for list-level throughput hints where the UI provides them. Open a pipeline to work in detail.
• Create — New pipelines are created from the Pipelines (and sometimes Overview) Create pipeline entry points. We support three flows: Build new (define from scratch), Migrate from Google SecOps Forwarder (upload a .conf and Praxis recreates the equivalent native pipeline), and Google SecOps Data Processing (managed runtime — no collector to install). See Create a Pipeline for the field-by-field walkthrough.
• Design — Visual builder for nodes and edges, draft edits, and actions such as validate and publish. Publishing a version rolls configuration out to collectors assigned to that pipeline (see your UI notifications for how many collectors were targeted). Some destinations (and related nodes) are tied to a saved integration record: the UI may block inline edits or removal for those nodes so configuration stays consistent with the integration your administrators maintain—update the integration (or ask an admin), then adjust the pipeline within your permissions.
• Throughput — Time-series and related views for pipeline edge throughput in the selected range (when applicable).
• Resource usage — Collector-side resource signals for pipelines that run on agents (when applicable to your graph).
• Collectors — List of collectors attached to this pipeline, rollout status, install instructions for new collectors, and actions such as push config, upgrade, reload pipeline (recycle pipelines in-process, no process restart), and restart (full collector process restart) — subject to permissions.
• History — Version history of published pipeline configurations, with compare and (where supported) restore of an older version to the current draft.

If you work on Google SecOps data processing pipelines (Chronicle-aligned flows in the management plane), you may see a smaller set of tabs—often Design and History only—because those flows are not centered on on-agent throughput the same way native collector pipelines are.

Collectors

Collectors is your inventory and control center for Praxis Collector agents: they connect to Praxis, pull pipeline configuration, and report metrics and host information.

• List — Table of collectors with hostname, IP, OS, assigned pipeline, version, pipeline version, config status, status, and last restarted, plus a rollout summary (for example how many of the selected or visible collectors are in a healthy “running” state) when the backend provides it.
• Bulk actions (when your role allows) — Upgrade, restart (full collector process restart), reload pipeline (in-process recycle of the otelcol pipelines without a full process restart — faster and lower-impact when you only need to pick up new configuration), push config (re-apply the latest published pipeline to selected collectors), and delete (with confirmation). A reload reports per-collector outcomes: ok, wedged (pipelines reloaded but the previous instance may need a full restart to clear), error, collector offline, or timeout if the collector never reports back.
• Collector detail — Open a single collector to see:
  • Overview — Identity, version, host metadata, CPU/memory/network/disk as reported, and related attributes.
  • Metrics — Collector and pipeline metrics over a selectable time range, with auto-refresh options where available.
  • Pipeline — A read-only view of the pipeline graph this collector is assigned to, so you can confirm configuration without opening the main pipeline editor.

Collectors are how the graphs you publish in Pipelines become real behavior in your networks and clusters. For install and sizing guidance, use the Installation overview.

See also

• Getting started — Adoption path and key terms
• API overview — Automating pipelines, collectors, and related resources
• Monitors — Metric-based alerts (separate from the Overview page’s traffic dashboards)