Trust Center

Security at OnCallReady — designed for the systems you can't afford to lose.

We operate on the principle of least privilege: collect only what's needed, audit everything, and keep customer data in customer infrastructure wherever possible. An agent that touches production must be held to a higher standard than a SaaS dashboard. We agree.

📋

Compliance & Certifications

We're honest about where we are. "In progress" is more useful to a buyer than a fake badge.

SOC 2 Type II

Audit underway. Auditor selected. Target completion Q2 2026. Controls across availability, security, and confidentiality trust service criteria.

In Progress

GDPR

Data minimization, purpose limitation, and processor agreement (DPA) available on request. EU data residency option available.

Aligned

HIPAA

Business Associate Agreement (BAA) available for Enterprise customers with healthcare data requirements. Reach out to discuss scope.

Enterprise

ISO 27001

On the roadmap following SOC 2 completion. Estimated program start Q3 2026.

Roadmap

Procurement reviews: Request our security questionnaire response (CAIQ/SIG), pen test summary, or sign an MNDA using the form at the bottom of this page. We respond within 1 business day.

🗄️

Data Handling

OnCallReady ingests only what's necessary to diagnose and resolve an incident. Here's exactly what we touch:

Ingested

Alert payloads

Webhook bodies from Datadog, Prometheus, Grafana, PagerDuty, or your custom source. Structured alert metadata — no raw log streams.

Ingested

Incident context

Outputs from runbook actions (command exit codes, service health check results). Minimal — enough to confirm resolution, not a full system snapshot.

In-memory only

Diagnostic telemetry

Intermediate diagnostic signals used during active resolution are processed in-memory and never written to disk or external storage.

Never collected

Customer application data

Database contents, user PII, API responses from your application, source code. OnCallReady has no reason to touch these and does not.

Retention

Standard: 90 days

Incident records, resolution logs, and runbook execution histories are retained for 90 days. Configurable on Enterprise.

Residency

US & EU

Choose where your incident data lives. US (AWS us-east-1) or EU (AWS eu-central-1). Configured at account provisioning.

Enterprise

BYOK

Bring Your Own Key: customer-managed encryption keys via AWS KMS or HashiCorp Vault. Available on Enterprise tier.

🔐

Infrastructure Security

At rest

AES-256 encryption

All stored data — incidents, runbooks, resolution records — encrypted at rest with AES-256. Database volumes encrypted via Neon's provider-managed keys by default; BYOK on Enterprise.

In transit

TLS 1.3

All traffic between clients and OnCallReady, and between OnCallReady and your infrastructure integrations, requires TLS 1.3. Older protocol versions rejected.

Secrets

HashiCorp Vault

Integration credentials and API keys are stored in Vault with short-lived dynamic secrets where supported. No long-lived credentials in environment variables or config files.

Network

Zero-trust internal

Internal services authenticate every request. No implicit trust by network position. Lateral movement requires explicit credential, not just VPC membership.

Production infrastructure runs on Render (web tier) with Neon PostgreSQL (data tier). Both providers maintain their own SOC 2 certifications. Network isolation between tenants is enforced at the database connection pool level.

🔑

Access Controls

Enterprise

SSO / SAML

Okta, Azure AD, and Google Workspace supported via SAML 2.0. OIDC on roadmap. SSO enforcement (no password fallback) available on Enterprise.

Enterprise

SCIM provisioning

Automated user lifecycle via SCIM 2.0. Offboard a user in your IdP — they lose OnCallReady access within minutes, not days.

All tiers

RBAC

Owner, Admin, Operator, and Viewer roles. Runbook authoring and execution are separate permissions. You control who can approve remediation actions.

All tiers

MFA enforcement

TOTP-based MFA required on all accounts. Enterprise customers can enforce MFA org-wide and block non-SSO login paths.

Audit log streaming: Every action — runbook execution, incident state change, user login, credential rotation — is written to the immutable audit log. Enterprise customers can stream this to their SIEM (Splunk, Datadog, Elastic) via webhook or S3.

🤖

Agent Permissions Model

This section matters most if you're evaluating OnCallReady for production access. The short version: the agent does as little as possible, and you control what it's allowed to do.

1

Read-only by default

OnCallReady operates with read-only credentials by default. Alert ingestion, diagnostic checks (health endpoints, metric reads), and runbook matching require no write access to your systems.

2

Explicit runbook authorization for write actions

Remediation actions (restart a service, drain a queue, rotate a credential) require a runbook that explicitly names the action class, the target resource, and the credential scope. No runbook = no action. You author the runbooks; we execute them.

3

Approval policies (optional)

Any action class can be gated behind a human approval step. Configure a Slack approval flow or email gate for destructive actions (database restarts, cache flushes). OnCallReady waits for the approval before proceeding. Timeout = escalate, not auto-approve.

4

Credential scoping

Each integration uses a dedicated service account with the minimum IAM role required by its runbooks. The Kubernetes integration used for "scale deployment" cannot be used to modify RBAC policies. Scopes are not inherited across runbooks.

5

Full audit trail

Every action taken (or rejected) by the agent is written to the immutable incident event log with timestamp, credential used, stdout/stderr, and exit code. You can replay exactly what happened and why.

🚨

Our Incident Response

How OnCallReady handles security incidents affecting the OnCallReady platform itself:

🔗

Sub-processors

These vendors may process customer data as part of delivering the OnCallReady service. We updated this table when sub-processors change; customers on Enterprise receive advance notice of additions.

Vendor Purpose Data accessed Region Compliance
Render Web application hosting Application code, request logs, environment variables US (Oregon) SOC 2 Type II
Neon PostgreSQL database Incidents, runbooks, resolutions, user records US / EU (configurable) SOC 2 Type II
Postmark (ActiveCampaign) Transactional email Recipient email address, email content US SOC 2 Type II
Stripe Payment processing Billing info, payment method (Stripe-hosted, not stored by OnCallReady) US PCI DSS Level 1, SOC 2
Cloudflare R2 File storage User-uploaded assets, report attachments US SOC 2 Type II, ISO 27001
OpenAI (via Polsia proxy) AI-assisted diagnostics (when enabled) Alert payloads passed to models for triage assistance US SOC 2 Type II
🛡️

Vulnerability Disclosure

We operate a responsible disclosure program. If you find a security issue in OnCallReady, please tell us before anyone else.

Request security documentation

Need our security questionnaire response (CAIQ / SIG), pen test executive summary, or an MNDA before your procurement review? Fill this in and we'll respond within 1 business day.