Operations platform
Observability that does not
bill you for cardinality
Atmora Signal ingests OpenTelemetry metrics, logs and traces into a single store with a hot window and a cheap aggregate tier. Alerts fire on service-level objective burn rate rather than instantaneous thresholds, which is why on-call engineers stop muting the channel.
- Cloud
- On-premise
- Hybrid
- Private cloud
Overview
Metrics, logs and traces on one engine, priced per host, with cardinality budgets per team.
Atmora Signal is observability for the team that owns the estate. Telemetry arrives over OTLP from any instrumented service, an optional host agent adds process, disk and eBPF network detail, and everything lands on one storage engine so a trace, the logs it produced and the host metrics underneath it are one query away from each other. Pricing is per monitored host with a cardinality budget per team, which is a deliberate stance: the usual failure in observability is not missing data, it is an invoice growing faster than the estate and an alert channel nobody reads.
Signal is built for platform and site reliability engineers, not for business reporting. It will not replace a BI tool, does not model revenue and has no opinion about your funnel. Estates below roughly twenty hosts are usually better served by their cloud provider's native monitoring, and we will say so during evaluation. Signal starts paying for itself with several clusters, a mixed cloud and on-premise footprint, and an on-call rota that needs its pages to be trustworthy at three in the morning.
- 11 minutes
- Median time to restore for paged incidents at Corvus Logistics
- -43%
- Observability spend after cardinality budgets and tiered storage
- 94%
- Pages acted on rather than silenced
- 1.2 million
- Active time series carried per monitored host
Modules
What ships in the box
- Metric store and query engine
- Log pipeline with parsing and routing
- Distributed tracing and service map
- SLOs, alerting and on-call schedules
- Synthetic and uptime checks
- Cost, capacity and cardinality control
Integrates with
- OpenTelemetry
- Kubernetes
- AWS CloudWatch
- PagerDuty
- ServiceNow
- Slack
- Terraform
- Okta
Capabilities
What it does
OpenTelemetry ingest, no proprietary agent required
Point an OTLP exporter at the collector and metrics, logs and traces land in the same store. The optional host agent adds process, disk and eBPF network detail where you want it.
Cardinality guardrails before the invoice
High-cardinality labels are detected at ingest, attributed to the team that emitted them, and dropped or aggregated by rule instead of being discovered three weeks later on a bill.
Tiered storage with a hot window you choose
Raw samples stay hot for fifteen days by default, then roll into five-minute aggregates on object storage. A query spanning both tiers is written the same way as any other query.
Burn-rate alerting on service-level objectives
Define objectives on latency, availability or a custom ratio. Pages fire on multi-window burn rate rather than a single threshold breach, so one bad minute does not wake anybody up.
Alert grouping and sane repeat behaviour
Related alerts collapse into one incident by service and suspected cause. Routing follows the on-call schedule, and repeat notifications back off instead of firing every sixty seconds.
Kubernetes discovery that survives a rollout
Targets are tracked by workload identity rather than pod address, so a deployment restart does not show up as a gap in the graph or as a fresh set of unrelated time series.
Trace to log in one click
Span identifiers are indexed against log lines at write time. Opening a slow span shows exactly the logs that request produced, on the same screen, without composing a second query.
Cost and capacity attributed per team
Spend maps to namespaces and tags, and headroom is shown against real request and limit values, so the argument about node counts is settled with numbers instead of instinct.
Built for
Where this is deployed
Industry
Banking & Financial Services
Payment rails, risk engines and core-adjacent systems built by people who have run them at 03:00 on a settlement night.
Industry
Telecom & Media
OSS, BSS and delivery platforms that keep working after launch night.
Industry
Retail & E-commerce
Storefronts, stock ledgers and fulfilment logic that hold together on your busiest trading day.
Industry
Public Sector
Digital services that work for everyone who has to use them, including on a five-year-old phone.
Industry
Healthcare & Life Sciences
Clinical and regulated software built to pass an audit, not just a demo.
Industry
Logistics & Supply Chain
Visibility, customs filing and cost-to-serve systems for operators who move real freight.
FAQ
Questions we get asked
- How is Signal priced, and what stops the bill from surprising us?
Per monitored host, with a per-team monthly budget for active time series and log volume. As a team approaches its budget the owning engineers are notified and new high-cardinality labels are dropped by rule, so an overspend becomes a conversation rather than an invoice.
- Can we run it fully air-gapped?
Yes. The collector, storage, query and alerting components ship as containers with no outbound calls required, and updates arrive as a signed bundle. Synthetic checks from external regions are the only feature that needs internet access, and they are optional.
- Will it replace our existing dashboards?
It reads Prometheus-style queries and imports Grafana dashboard JSON, so most panels come across intact. If your team is happy with a Grafana front end, keep it and use Signal as the store and alerting layer underneath; there is no requirement to adopt our interface.
Related
More Operations platforms
Operations
Inventory Management
Multi-warehouse stock, batch traceability and counting on handheld scanners.
Operations
Visitor Management
Visitor, contractor and delivery access control for multi-site campuses.
Operations
Booking Platform
Slots, resources and reschedules for organisations where an empty hour costs real money.