Skip to content
Technology

The system being up on payroll day is not a coincidence

On the last day of the month everyone logs in at once, timesheets close and payroll is calculated. The infrastructure was built for that one day: independent services, automatic scaling and a written service level commitment.

99.9%Monthly uptime commitment
AES-256Encryption at rest
HourlyDatabase backups
ISO 27001Certified management system
Architecture

Six decisions, all from the same principle

The system was designed not to “never break” but to “recover when it does”. Each heading below is an application of that principle.

One module does not bring down another

Payroll, leave, attendance, ATS and performance run as separate services. Updating one module, or one straining under load, does not stop the others.

Independent containers, REST between services, event-based messaging

In the cloud, with nothing to install

The application and data layers run on AWS; there are no servers, licences or maintenance windows on your side. Capacity grows automatically during busy periods.

AWS Europe (Frankfurt / eu-central-1), multi-availability-zone deployment

Data encrypted in transit and at rest

Traffic from the browser to the server and records sitting on disk are both encrypted. Encryption keys are held in a service separate from the application.

TLS 1.3 in transit, AES-256 at rest, key management via AWS KMS

Customer data never mixes

Each company’s data sits in a logically separated schema. A query reaching another company’s records by accident is prevented architecturally.

Tenant-based isolation, role-based permissions, full audit log

Taking backups is not enough — you have to restore

Backups are taken regularly and kept for thirty days. What matters more is rehearsing the restore: the disaster recovery scenario is tested on a schedule.

Thirty-day retention, monthly recovery drill, point-in-time restore

Not a closed box

It talks to your existing ERP, accounting and payroll systems. Where there is no ready-made integration, you connect over API and webhooks.

REST API, webhook triggers, Swagger-based documentation

Service level

The commitments written into the contract

Every line in this table appears in the service agreement in the same form. What happens when one is missed is defined there too.

HR&Tomorrow service level commitments
MetricCommitment
Monthly uptimeExcluding planned maintenance99.9%
Planned downtime ceilingOutside business hours4 hours / month
Support response — P1 (system down)Always-on channel30 minutes
Support response — P2 (function broken)Within business hours4 hours
Database backup frequency30-day retentionHourly
RTO — time to bring the service backIn a disaster scenario4 hours
RPO — accepted data lossRestore to the last backup point1 hour

Planned maintenance is announced at least five working days in advance and carried out outside business hours. No planned maintenance takes place on payroll closing days.

Security

Not one door, but four

Getting past one layer does not invalidate the others. Four sets of controls, ordered from the outside in.

  1. Network edge

    Before a request reaches the application

    A web application firewall cuts known attack patterns, and DDoS protection stops volumetric attacks before they reach the application.

  2. Identity

    Before you are confirmed as who you say you are

    Role-based permissions, session control, and multi-step verification for administrator accounts. Access can be narrowed at location and module level.

  3. Application

    Services do not trust each other

    Every call between services is separately verified. Every action — who, when, which record — is written to an auditable log.

  4. Data

    Everything written to disk

    Encrypted storage, tenant-based isolation, and a defined deletion procedure for data whose retention period has expired. Backups carry the same encryption.

Certifications and compliance

We document what we claim

Audit reports and certificate copies are all shared during the purchasing process.

ISO 27001

Information security management system certification. Evidence that processes are documented, responsibilities assigned and auditing regular.

SOC 2 Type I

A report in which the security, availability and confidentiality principles are examined by an independent auditor.

KVKK compliance

Obligations as a data processor under Turkish data protection law: a category-based processing inventory, a retention and destruction policy, and a channel for requests. Details

TLS 1.3

All client traffic is encrypted with the current protocol. Certificate management runs on automatic renewal.

Let’s answer your IT team’s questions directly

We run technical briefings on architecture, integration and data security. An engineer from our side joins the call too.