onshore
Products / Stop the leaks

Your logs can leave. The IDs in them don't.

Application logs are full of account numbers, and they ship abroad all day. The log processor sits between your apps and your log vendor and tokenizes what it finds on the way through.

On the way to your log vendor
level=error msg="refund failed" iban=DE89370400440532013000
level=info user=cust_1042 aadhaar=234567890124
level=warn ssn=123-45-6789
The problem

A single debug line can send thousands of customer identifiers to a vendor in another country, and nobody on the team decided that it should.

How it works
01

Change one endpoint

Point your OpenTelemetry Collector, Fluent Bit or Vector output at the processor.

02

Watch first

Run in Watch only to see what would be caught, with your logs unchanged.

03

Turn on protection

Identifiers become tokens. Everything else reaches your vendor as it was.

In code

OpenTelemetry Collector

exporters:
  otlphttp/onshore:
    endpoint: https://logs.gb.onshorevault.com/otlp/${env:ONSHORE_PIPELINE}
    headers:
      X-RV-Pipeline-Key: ${env:ONSHORE_PIPELINE_KEY}

service:
  pipelines:
    logs:
      exporters: [otlphttp/onshore]
What's included

Three front-ends, one engine

OpenTelemetry, Fluent Bit and Vector all use the same detection rules.

No code changes

Your application keeps logging as it does today.

Debugging still works

Deterministic tokens mean you can still follow one customer through your logs.

Works with OTLP

HTTP and gRPC for logs, using your existing pipeline.

Traces and metrics too

OTLP traces and metrics can pass through the same pipeline.

A key per pipeline

Each pipeline has its own ingest key, and a wrong key is refused.

Who uses this

Start with one leak.

Tell us which flow worries you most. We'll show you what it would take to keep it onshore.