Air-gapped deployment

Every Tracecat feature works in self-hosted and air-gapped deployments. We ship a feature to Tracecat Cloud only when it can also run on-premises.

Tracecat supports fully air-gapped operation on production Kubernetes. Use internet access during preparation to stage every required artifact, then run Tracecat with zero internet connectivity.

Reference architecture

The inference service can run inside Kubernetes or on an adjacent private GPU cluster. Every runtime connection stays on customer-controlled networks.

Prepare while connected

Mirror the Tracecat OCI chart and every referenced container image into an internal OCI registry.
Stage the inference image, model weights, tokenizer, and model configuration in internal repositories.
Provision internal PostgreSQL, Redis, S3-compatible storage, Temporal, identity, DNS, TLS, secrets, telemetry, and integration endpoints. Validate image pulls and model discovery before removing egress.

No runtime workflow, integration, telemetry exporter, identity provider, or package dependency may require a public endpoint.

Choose the model route

Configure the route per agent, including subagents.

Route	Choose it when	Request path
Direct passthrough	You already operate an OpenAI-compatible endpoint and want the shortest path or per-agent routing.	Agent executor → internal vLLM or Ollama
Bundled LiteLLM	You want centralized model aliases, credentials, request normalization, or multiple inference backends.	Agent executor → bundled LiteLLM → internal inference

Use vLLM as the distributed production option; confirm its model-specific tool parser. Use Ollama as the simpler quantized or single-server option; review its GPU, context, model import, and tool-calling guidance.

Open model reference

Reviewed September 2026. Each DGX Spark provides 128 GB of unified memory. Static weight figures exclude KV cache, serving overhead, and concurrency. Calculated figures use two bytes per BF16 parameter, one byte per FP8 parameter, and half a byte per FP4 parameter.

Model	Compute against 2–3 DGX Spark nodes	Size and precision	Tool calling	Context: input / output	Defensive cyber evidence
Gemma 4 26B A4B	Published load: 57.7 GB BF16, 28.8 GB SFP8, 14.4 GB Q4	25.2B total / 3.8B active	Native function calling; parser not listed in vLLM docs	256K / unavailable	Function calling published; defensive SOC result unavailable
Gemma 4 31B	Published load: 69.9 GB BF16, 34.9 GB SFP8, 17.5 GB Q4	30.7B dense	Native function calling; parser not listed in vLLM docs	256K / unavailable	Function calling published; defensive SOC result unavailable
Ministral 3 14B	Published: under 32 GB BF16 or 24 GB quantized	14B dense	Native function calling and JSON; vLLM `mistral` parser	256K / unavailable	No checkpoint-specific defensive SOC result found
Mistral Large 3	Calculated: ≈675 GB FP8; vendor reference uses 8×H200	675B total / 41B active, FP8	Native function calling and JSON; vLLM `mistral` parser	256K / unavailable	No checkpoint-specific defensive SOC result found
Qwen3-Next 80B A3B	Calculated: ≈160 GB BF16 or 80 GB FP8; model card uses TP4	80B total / 3B active	Tool use published; parser not listed in vLLM docs	262K native, 1.01M with YaRN / unavailable	Tool use published; defensive SOC result unavailable
Qwen3 235B A22B	Calculated: ≈470 GB BF16 or 235 GB FP8; vLLM example uses TP8	235B total / 22B active	Tool use published; parser not listed in vLLM docs	262K native, 1.01M with YaRN / unavailable	Tool use published; defensive SOC result unavailable
GLM-4.7-Flash	Calculated: ≈60 GB BF16 or 30 GB FP8	30B total / 3B active	Native; vLLM `glm47` parser	200K / unavailable	No checkpoint-specific defensive SOC result found
GLM-5.3-Flash	Calculated: ≈320 GB FP8 or 160 GB NVFP4; two-Spark NVFP4 reference	320B total / 18B active, FP8/MTP	Native; vLLM `glm47` parser	1M / unavailable	AutomationBench published; defensive SOC result unavailable
Kimi-Linear 48B A3B	Calculated: ≈96 GB BF16 or 48 GB FP8	48B total / 3B active	OpenAI-compatible serving; native tool-use evidence unavailable	1M / unavailable	No checkpoint-specific defensive SOC result found
Kimi K3	vLLM reference: ≈1.68 TB across 16 GPUs; no 2–3 Spark reference	2.8T total / 104B active, MXFP4 weights	vLLM `kimi_k3` parser; provider notes format caveats	1.05M / unavailable	Toolathlon and AutomationBench published; defensive SOC result unavailable

Cyber evidence means checkpoint-specific evaluation for defensive alert triage, investigation, structured tool use, or workflow generation. General reasoning, coding, and offensive-security scores are not substitutes. Track emerging defensive evaluations such as SIABench and CyberSOCEval. Treat static weight fit as a starting point, not a capacity guarantee. Validate target context length, KV cache, concurrency, latency, and serving overhead with your own workload.

See Kubernetes for component and resource requirements.
See Self-hosted architecture for service boundaries and request paths.
See Security architecture for the threat model and trusted execution boundaries.
See TLS and certificates for internal certificate authorities and ingress TLS.
See Custom LLM providers for passthrough configuration and model discovery.

Overview

Building automations

Resolve cases

Building agents

Building custom integrations

Self-hosting

Manage platform

Authentication

Security

Reference architecture

Prepare while connected

Choose the model route

Open model reference

​Reference architecture

​Prepare while connected

​Choose the model route

​Open model reference

​Related pages

Reference architecture

Prepare while connected

Choose the model route

Open model reference

Related pages