Bad developer UX? Look out for breakages, burnout, and "biological single points of failure"
"All of this hasn't existed for the CIO or the CTO, and genuinely they've been begging for a platform with the same ability to report back to executives to show the value of their engineers..."
Nick Durkin, Field CTO of Harness, takes resilience seriously. He previously acted as lead architect on critical infrastructure for the for the US government, including the Department of Homeland Security’s financial verification platform. Durkin is big on fail-safes, and wishes software companies took the human element of them more seriously.
Too often, he says to The Stack, "the way companies work is that they rely extensively on their best developers -- and when you do that, you are setting yourself up for biological single points of failure."
What Durkin is talking about is retention – and burnout.
Developers spend so much of their day doing "the most boring part of their jobs," Durkin says. His criticism is backed up by numbers -- in a 2022 report, DevOps, an industry magazine, reported that 83% of software developers felt burned-out; cue serial job-hopping and low productivity. (Developers romancing over the idea of becoming farmers on forums like Reddit is something of an industry cliché at this point.)
And when your top talent leaves, frustrated and overworked, it can have damaging consequences.
No tool is going to magically right the myriad ills that can drive a talented developer out of their workplace in search of more money, more creativity, less management, fewer meetings, or perhaps, goats. But Durkin says Harness wants to make the intolerable tolerable and the tolerable, fun.
"What our goal is with Harness, is to be that 'safety harness', if you will, and take away the worst parts of a developers work and automate it."
The idea is speed up test cycles, automate governance of build and deploy processes while also providing developers with faster feedback loops, explains Durkin. It isn't just about automating though- it's also about that key fail-safe and roll back mechanism for when something goes south.
"The biggest case in point for this is the Knight Capital incident of 2012..." Durkin trails off, referring to the devastating $440 million software error, where a glitch within a new trading software sent the company's computers on a buying spree, accidentally purchasing 150 different stocks at a total cost of around $7 billion, all in the first hour of trading.
These sort of errors can bankrupt firms, and destroy consumer confidence, explains Durkin -- and he suggests feature flagging systems and emergency roll backs to all clients.
See also: A $1bn fat finger helps drive Citi's $13.6bn IT spending plans
"You want to get an alert when there's any error within the new feature you've deployed -- not after its cost you thousands," this is a lesson learnt from experience for Durkin and his team at Harness.
"We actually built the feature flagging tool after our last outage -- we used to use internal database toggles like most companies would do -- which would give customers region-specific pages. [Then we suffered an] outage because we deployed ourselves and didn't realise that a developer had flipped a flag which had no governance or auditability..."
"We rolled back and it still wasn't fixed. We spent three hours troubleshooting the problem and couldn't find what it was. It turned out it was just that a developer had flipped a flag [feature flags allow you to enable or disable a feature without modifying the source code] and at that point we said, no more of this. We learnt that we didn't have the audit abilites; couldn't see who flipped the flag or when they flipped a flag or prevent the flag from being flipped... so that is where the new feature came in!"
Another real issue for not just developers but CIOs is that many organisations don't pay them the blindest bit of attention until something breaks. A recent debacle at Waitrose is a case in point. A bungled IT update caused errors within supply chain and delivery orders for the grocery giant, leaving shelves across the country empty before Bank Holiday weekend- contributing to a 3% in overall sales for the loss-stricken grocer.
No one realises the developer is in the picture till something goes wrong, Durkin laughs. Having been a software developer for over two decades himself, the pain of unacknowledged labour is personal.
"There are no ways for a CIO to brag, when it comes to metrics," he explains in regards to the ways in which software developers and innovation teams can go feeling unrecognised in many organisations.
"Every other executive who sits in a boardroom has a platform -- whether it's Salesforce or it's ServiceNow or it's Workday- they all have their own platform. But the CIO doesn't have that,"
"There is no great way for us to report on how efficient our engineering organisations are. There's no good way to understand things like the DORA metrics easily or things like how often are we committing code, how often is it breaking? And how quickly can we get it back if it does break," he explains.
"All of this hasn't existed for the CIO or the CTO, and genuinely they've been begging for a platform with the same ability to report back to executives to show the value of their engineers -- that's the platform we're building.
Appreciation, along with being empowered to innovate and not just maintain and toil has is key to retaining top tech talent, concludes Durkin.
"There are many people out there to poach your best talent. The way that I look at it, I want every one of my engineers to be getting LinkedIn requests for job offers every day. Then I know I have the right engineers, the good ones – but what I also want is that we provide enough value and we we make their lives easy enough and efficient enough, that they don't even open that InMail."
