Systems Reliability at Delimited

Our philosophy is that the reliability of your infrastructure should be such that you can completely forget about it, except to occasionally think, “I put that in place years ago and it’s still going.” Wouldn’t it be great if all of our foundational systems hit that target?

We build Delimited with the above philosophy. Below are some of the aspects of the Delimited infrastructure that mean you can trust it for high reliability.

Foundational Infrastructure

Delimited is built using modern and reliable cloud infrastructure with auto-scaling capabilities to meet service demand. The service is self-healing, with automated rollbacks for any deployments deemed to be unsatisfactory. Data is kept in scalable, secure, private, and highly-available cloud services. Any single point of failures are eliminated or are actively being mitigated.

Development Process

Development on Delimited is done using industry best practices around DevOps, such as infrastructure-as-code, code reviews, automated continuous integration, continuous deployment, and continuous delivery. This combines resiliency with an incredible amount of flexibility and agility in delivering features and bug fixes.

All production secrets are kept in encrypted stores and are never maintained in source control.

Testing

Automated tests are used in all stages of the development and operational lifecycle. Testing provides assurance that the system is working as designed, and a well-tested system is a necessity for any reliable system. Test types include:

• Unit testing for individual software components
• Integration testing to test that system components work well together
• Regression testing to reduce recurrence of previously discovered bugs
• Smoke testing to provide assurance of success after deployments
• User interface testing to check that the system is working as design for users

Monitoring

Continuous production monitoring for Delimited shows availability, latency, and telemetry of underlying hardware (such as CPU and memory usage), among other things. Monitoring is done 24/7 through automated systems and immediately informs our operations engineers if anomalies are detected that warrant investigation. System components have “health checks,” which are monitored as part of the self-healing architecture.

Outages and Remediation

While we work extremely hard to prevent any production outages, they may occasionally occur. During a production outage we commit to communicating transparently with our customers about the outage and expected timeframes for restoration of services. After any production outage, the Delimited team investigates and performs a full root cause analysis, with recommendations for changes to development and operational processes or systems so that the outage does not happen again.

Data Lifecycle

All data is subject to strict privacy and operational constraints, including encryption at rest and in transit and automated data lifecycle management. For example, Delimited completely deletes – NOT soft deletes – customer-uploaded data after a pre-defined length of time. Data is kept for long enough to allow administrators to troubleshoot issues and download original copies, as needed (currently 30 days at time of writing), and then is completely erased through automated systems.

Threat Detection

Audit logs are kept across the Delimited infrastructure. Automated threat and intrusion detection software is use to detect anomalous or suspicious behavior. Any findings are promptly addressed.

Delimited is Reliable

While each of the above mentioned categories is important, together they form the core of the systems reliability of Delimited, and they make Delimited a reliable product that we’re proud of. We recognize that our success is based on our customer’s trust, and that trust is maintained through reliability and security of our infrastructure and product.

If you’d like to give Delimited a try, sign up for a free trial on our Pricing Page.