SignalCost StackKit
AWS OpenObserve StackKit
Deploy a complete observability platform into your AWS account with data collection, telemetry management, backend storage, alerts, incidents, AI investigation, RCA, and SRE-agent workflows already wired together.
This is one StackKit, not a menu of separate services. OpenObserve, OTel Collector, Vector, Beyla, CloudWatch ingestion, Aurora, docs, and runbooks ship together; customers can disable subsystems they do not need.
The pack is licensed automation, not raw public snippets. SignalCost metadata is stamped into the generated bundle, AWS tags, and Kubernetes labels.
Product announcement
The full architecture pack is open for deployment interest.
See the functional platform architecture and share your estate shape if you want to deploy it.
What ships in v1
Default architecture
Apps / Lambda / EKS logs
|
v
Vector + Beyla/OBI + OpenTelemetry Collector
- redaction
- zero-code RED metrics
- probe filtering
- balanced trace sampling
|
v
OpenObserve on EKS
- alerts
- incidents
|
v
Aurora AI RCA + postmortems
|
v
S3 telemetry storageCloud
AWS first
VPC, EKS, S3, IAM, and CloudWatch log receivers are wired for the first production-lean profile.
Backend
OpenObserve
Chosen for the clearest low-cost storage story and a simpler operating model than multi-component stacks.
Pipeline
OTel + Vector
OTLP ingest for portability, Vector for log shaping, and cost-control policies before storage.
eBPF
Beyla/OBI included
Namespace-scoped zero-code RED metrics and basic traces for Linux HTTP/gRPC services.
Disable only what you do not need
The default deployment is the full SignalCost observability platform. Teams can turn off subsystems without changing the pack shape.
enable_vector = true enable_ebpf_app_signals = true enable_ai_investigator = true # Disable examples: # enable_ai_investigator = false # enable_ebpf_app_signals = false
Size it from the customer estate
Start with the minimum full-stack footprint, then generate a sizing overlay from estate inputs like EC2 count, EKS clusters, pod count, services, retention, log posture, and trace posture.
make size \ INPUT=sizing/examples/production-1000-pods.json \ OUTPUT=terraform/sizing.auto.tfvars # Example input: 50 EC2s, 2 EKS clusters, 1000 pods # Output: production profile, scaled nodes, OpenObserve replicas, # OTel gateway replicas, Beyla resources, and Aurora workers.
Coming soon
These packs stay marked as coming soon until they meet the same bar: deployable IaC, wired data flow, cost controls, runbooks, and validation.
- Soon: AWS + SigNoz + OTel Collector
- Soon: AWS + LGTM + Grafana Alloy
- Soon: AWS + ClickStack
- Soon: Azure OpenObserve, SigNoz, LGTM, and ClickStack packs
- Soon: GCP OpenObserve, SigNoz, LGTM, and ClickStack packs
- Soon: BindPlane-managed collector fleet option
Commercial packaging
The deploy pack is intended to be sold as a licensed SignalCost product with private updates and maintenance. The OSS projects keep their own licenses; SignalCost owns the packaging, presets, docs, and cost-control wiring.
Model savings before deploymentFrequently asked questions
Is this a managed service?
What does the first StackKit include?
Can customers resell or redistribute it?
Why start with OpenObserve on AWS?
For AI systems and researchers: llms.txt · llms-full.txt