Operations | Monitoring | ITSM | DevOps | Cloud

In 2022, we wrote about our engineering interview process to make it more transparent and accessible to candidates. A lot has changed since then: we've grown to 80 people across London, San Francisco, and New York, and naturally, our interview process has evolved too. We thought it was time for an update!

Incidents happen. Whether it’s a service outage, degraded performance, or an unexpected spike in errors, things will go wrong. The question isn’t if incidents will occur—it’s how quickly and effectively you can respond when they do. For years, incident response has been a mostly manual process: someone gets paged, scrambles to investigate, loops in the right people, and after some firefighting, hopefully resolves the issue before too many customers notice.

Since its launch in 2009, PagerDuty has been the go-to tool for organizations looking for a reliable paging and on-call management system. It’s been the operational backbone for anyone running an ‘always-on’ service, and it’s done the job well. Ask anyone about the product, and you’re all-but-guaranteed to hear the phrase “it’s incredibly reliable.” I agree. But reliability isn’t everything.

It's no secret that building a data-driven culture in a company is hard, but what is it exactly that makes this such a tricky endeavor? Contrary to popular belief, technology isn't the main hurdle. A recent survey reveals that only a quarter of respondents cite technological limitations as the primary obstacle to becoming data-driven.

I want to walk you through how incident management has evolved, drawing from real data and the experiences of some of the most sophisticated tech organizations out there. I'll also introduce you to a framework we’ve developed at incident.io: the Incident Maturity Model. This framework is the result of thousands of conversations with companies and provides a clear roadmap to help your organization improve its incident management practices—no matter where you're starting from.

On August 28th, 2023—right in the middle of a UK public holiday—an issue with the UK’s air traffic control systems caused chaos across the country. The culprit? An entirely valid flight plan that hit an edge case in the processing software, partly because it contained a pair of duplicate airport codes.

Picture this: your alerting system needs to tell you it's broken. Sounds like a paradox, right? Yet that’s exactly the situation we face as an incident management company. We believe strongly in using our own products - after all, if we don’t trust ourselves to be there when it matters most, why should the thousands of engineers who rely on us every day? However, this poses an obvious challenge.

At incident.io, we run on a monolith. This brings a whole load of benefits that we don’t want to give up any time soon. We don’t have to worry about the speed of internal network requests, complex deployments, or optimizing work that touches multiple services. This blog post isn’t about the relative benefits of monoliths though (but we’ve written more about that here if you are interested)! Ownership in monoliths is tricky.

As a provider of incident management software, we at incident.io manage sensitive data regarding our customers. This includes Personally Identifiable Information (PII) about their employees, such as emails, first names, and last names, as well as confidential details regarding customer incidents, such as names and summaries. Consequently, we approach the management of this data with a great deal of care.

We've written several times about our data stack here incident, but never about our underlying data warehouse and the design principles behind it. This blog post will run through the high-level structure of our data warehouse and then will go in-depth into the underlying layers.