Security & trust
We hold your customers' requests. Here is exactly how.
Flintwake sees traffic, so the interesting questions are what we capture, where it lives, who can read it and how quickly it disappears. This page answers those four questions without marketing language.
- 1.8ms medianAdded latencyMeasured at the collector across the demo fleet
- Fails openFailure modeIf capture cannot write, your request still answers
- TLS 1.3TransportCollector to region, and region to console
- AES-256At restObject store and search index
Data
Four answers, in order of how often we are asked.
Redaction happens before capture
Capture rules are evaluated inside your own network boundary, at the collector. A field matched by a redaction rule is replaced with a typed placeholder before the wake is serialised, so the original value is never written and never transmitted.
- JSONPath rules, versioned and reviewable in Git
- Header allowlists rather than denylists
- Whole-body exclusion available per endpoint
Residency is a project setting
Each project is pinned to one processing region — US, EU or AP. Wakes captured for that project are written, indexed and served from that region only. Console metadata is replicated; captured payloads are not.
- us-east-1, eu-central-1 and ap-southeast-2
- No cross-region payload replication
- Region is fixed at project creation
Retention is short by default
Seven days on Starter, thirty on Growth, ninety on Scale — and you can set it lower per environment. Deletion is a hard delete from the index and the object store, not a flag.
- Per-environment retention override
- Hard delete within 24 hours of expiry
- Deletion receipts available on request
Access is scoped and logged
Console roles map to what someone can see: metrics without bodies, bodies without keys, or full administration. Every read of a captured body is written to the audit log with the actor, the wake and the time.
- Four built-in roles plus custom scopes
- Body reads are audited individually
- SSO and SCIM on Scale and above
Redaction
The rule file is the contract.
Rules live in your repository, go through your review process and are applied at the collector. A field matched here never becomes a byte we hold.
Header handling is an allowlist, not a denylist: anything you have not explicitly permitted is dropped before capture, so a new header added by a proxy cannot quietly start being recorded.
{
"version": 7,
"capture": [
{ "match": "/v2/**", "bodies": true },
{ "match": "/v2/identity/**", "bodies": false },
{ "match": "/v2/health", "capture": false }
],
"redact": [
"$.card.number",
"$.card.cvc",
"$.customer.email",
"$.customer.national_id",
"$.*.tax_id"
],
"headers": {
"allow": ["content-type", "user-agent", "x-request-id"]
},
"region": "eu-central-1",
"retention_days": 30
}Practices
How we run the service.
Engineering
- Two-person review on every change to capture or redaction
- Dependency and container scanning on every build
- Infrastructure defined in code and reviewed in the same flow
Access control
- Single sign-on with hardware keys for every employee
- Production access is time-boxed and approved per session
- No standing access to captured payloads
Operations
- On-call rota with a documented escalation path
- Quarterly restore tests from backup
- Public status page and incident write-ups within five working days
Third parties
- Subprocessor list published and versioned
- Notice before a new subprocessor handles captured data
- Annual review of every vendor with data access
Flintwake is a fictional product built as a website template. This page describes practices only — it makes no certification, audit or regulatory compliance claim, and none should be inferred. Replace every statement here with your own verified position before you publish.
Questions
The ones security teams actually ask.
Only if you ask for them, and only after your redaction rules have run. Body capture can be disabled entirely, per endpoint or per environment.
Send us your questionnaire.
Security review is part of the evaluation, not a hurdle after it. Write to security@flintwake.com and a human who can answer will reply.