Skip to content

Applications · Infrastructure · Network · Security

Observability and security, from code to cable.

One OpenTelemetry-native platform where the teams that keep your systems running and the team that keeps them safe work from the same data. Every workspace, on every plan, gets its own database.

Free plan: 1 workspace, 3 users, 1 GB of telemetry a day and 7-day retention.

Two teams, one estate

Operations and security, working from the same data

Most teams point separate tools at the same hosts, clusters, cloud accounts, firewalls and switches. Tracix gives operations and security one collector, one isolated database and one alerting engine, so a signal on one side is context on the other.

Shared by both teams

  • Tracix Collector

    OpenTelemetry distribution with fleet management

  • Isolated workspaces

    Own namespace, database and DNS name

  • Monitors & alerting

    One engine for a slow service, a failing port or a security match

  • Anomaly detection

    Seasonal baselines that score what is unusual

  • Problems & cases

    Related alerts in one problem, the response in one case

  • APIs & as code

    Monitors, dashboards and detections as code

Two stories, one estate

Whether something breaks or someone breaks in

Both stories run through the same Kubernetes pool. Tracix keeps the evidence together from the first signal to the last approval, so nobody has to rebuild the story in a chat thread.

When something breaks

Follow a problem from the code to the cable

  1. 01 · DetectThe checkout SLO starts burningA burn-rate monitor pages on-call before customers notice.SLOs · On-call
  2. 02 · CorrelateSignals become one problemThe latency spike, the port errors and the config change land in one problem, not three tickets.Problems
  3. 03 · ExplainBrain names the likely causeAn MTU mismatch from change 4821, with links to the trace, the port and the diff.Brain
  4. 04 · FixA reviewed rollback goes outNCM drafts the rollback, checks it against policy, and a collector on your network applies it after approval.Configuration management
  5. 05 · VerifyRecovery is measured, not assumedSynthetic checks and the SLO confirm the fix, and the case closes with a full timeline.Synthetics · Cases

When someone breaks in

Follow an intrusion from the first event to an approved response

  1. 01 · DetectA SigmaHQ rule catches a reverse shellIt fires on pool-b-node-3, from Linux endpoint events the collector already sends.Threat detection
  2. 02 · TriageThe triage agent returns a verdictMalicious, 0.93 confidence, routed to the incident lane with its evidence cited.SOC triage agent
  3. 03 · ScopeAttack paths show what it could reachPick the host and see everything it reaches within six hops, with the evidence for each.Attack paths
  4. 04 · RespondContainment waits for an adminA proposal to contain the host through CrowdStrike runs only after an admin approves exactly what was proposed.Response & playbooks
  5. 05 · RecordEvidence that holds up laterThe case chains its evidence by hash, and every action lands on a ledger you can verify.Cases · SOC ledger
The outage ends in a case: rollback CR-1052 approved at 14:10, checkout back to normal at 14:16.
So does the intrusion: case 312 keeps its evidence hash-chained, and the custody chain verifies.

Detection rules come from the SigmaHQ project under the Detection Rule License 1.1. MITRE ATT&CK® is a registered trademark of The MITRE Corporation.

Anomaly detection

Machine learning that watches for both teams

Anomaly jobs learn the usual range of each series, with daily or weekly seasonality, and score what falls outside it from 0 to 100.

  • Metric jobs page on-call only when an anomaly lasts, and a gap in the data never counts as recovery.
  • Rare-value, population, log-pattern and fleet-outlier jobs raise leads: the value that never shows up, the host unlike its peers, the log line nobody has seen before.
  • Backtest a job over up to 30 days before you save it.
  • Every workspace runs its own anomaly worker.
How anomaly detection works

Tracix Collector

Install once. Both teams get their data.

The Tracix Collector is an OpenTelemetry Collector distribution with fleet management built in. Run it on Linux, macOS, Windows, Kubernetes or Docker, and the same agents send what operations and security both need.

  • Policies push configuration to groups of agents, and every change is a new, signed revision.
  • Linux process, file and network events feed threat detection, with no second agent to roll out.
  • The software inventory of running Linux apps, containers and pods is matched against OSV advisories, with KEV and EPSS beside each match.
  • Discovery finds SNMP devices across CIDR ranges and inventories AWS, Azure and GCP.
  • Standard OTLP in and out, so your existing SDKs keep working.
$ curl -fsSL https://charts.tracix.io/install.sh | sudo bash -s -- \  --endpoint https://agents-acme-prod-x7k2.app.tracix.io \  --token-file /root/tracix-token==> Tracix Collector — Linux install==> verified SHA-256 checksum==> starting tracix-collector.service==> enrolled; waiting for the policy configuration==> Installation complete!
Policy linux-servers · revision 14 · signed128 agents healthy

Isolated by design

Every workspace gets its own database. On every plan.

Each workspace runs in its own Kubernetes namespace with its own ClickHouse database and DNS name. Your telemetry never shares a database or a storage bucket with another customer's, and your lifecycle rules decide when it moves to cold storage or is deleted.

Read how isolation works
  • Single sign-on and SCIM

    Provision people from your identity provider and keep access in step with it.

  • Multi-factor authentication

    Authenticator apps with backup codes for every member.

  • Built-in and custom roles

    Owner, Admin and Member, plus custom roles across 52 permission surfaces.

  • Credential vault

    Device credentials stay in the vault. Agents get a scoped, short-lived grant per job.

  • Audit trail

    Policy, agent and configuration changes are recorded with who, what and when.

  • Data lifecycle

    Hot and cold storage tiers, scheduled deletion and legal holds.

  • Scoped API keys

    Give scripts and CI jobs only the access they need.

  • Custom domains

    Serve a workspace from a domain you own, with a managed certificate.

Brain

An analyst for both teams that asks before it acts

Brain answers questions about a slow service or a security alert in plain language, explains with links to the evidence, and turns a fix into a workflow that runs only after the right person approves it.

  • Ask

    Query logs, metrics, traces, devices and alerts in plain language.

  • Triage

    Turn on the SOC triage agent and each firing alert gets a verdict, a confidence and a lane, citing evidence.

  • Automate

    Build workflows and agents on a typed graph, with Auto-Fix proposals.

  • Govern

    Decide which agents can act, on what, and with whose approval.

Bring your own model: Anthropic, OpenAI, or an OpenAI-compatible endpoint you host, such as Ollama or vLLM.

Meet Brain

Integrations

Works with the tools both teams already run

238 data collection integrations ship with 146 dashboards and 125 log pipelines, and 1,152 action packs let workflows act in the tools your teams use. Upstream OpenTelemetry SDKs and the protocols your estate already speaks, from SNMP and IPFIX to Syslog, cover the rest.

Browse all integrations

See your own estate in Tracix

Start on the Free plan in a few minutes, or book a demo and we will walk your operations and security teams through your environment.