DevOps: Great Musical Collaborations or Generalist Looney Tunes?
DevOps is the promoter of business tech versions of strange but purposeful partnerships.
The music business is full of wonderful collaborations. Think Eminem and Rhianna with “Monster”, the Art of Noise teaming with perennial crooner Tom Jones to cover Prince’s “Kiss”, or even the late Freddie Mercury and Monserrat Cabelle belting out “Barcelona”. On the face of it these very different styles shouldn’t work together, but by some miracle of musical chemistry and maybe a slice of production wizardry, they just do—brilliantly! So much so that many a flagging career has been revived by an unexpected hit.
DevOps is the promoter of business tech versions of strange but purposeful partnerships. Dev rappers teaming with IT operations heavy rockers to inject quality into application releases. Operatic enterprise architects pairing with ambient style agilista’s to ensure sweet but supportable software compositions. Great combos, collaborating to deliver hit after business hit.
So why does the industry try to screw it all up? Insisting that in the name of DevOps, any one group of talented tech musos should become skilled in other areas.
Of course, it all sounds plausible. After all, shouldn’t building an ensemble of DevOps generalists become the catalyst for busting silos and removing entrenched thinking?
Unfortunately it’s not so clear-cut. So before handing out the “I’m a DevOps engineer—trust me!” and radically overhauling organizational structures, consider that “one-man band, you hum it I’ll play it generalists” can actually create as many problems as they attempt to solve.
Business metrics are what really counts, so DevOps tools and dashboards should include indicators that enable teams to focus on biggest problem everyone needs to solve—improving the customer experience.
— Peter Waterhouse, Senior Strategist, CA Technologies
Would You Ask a Guitar God to Play the Bagpipes?
Just as the music world wouldn’t have asked Hendrix to play the Sax or Pavarotti to sing soprano, the IT industry should recognize that tech pros are hard-wired to do certain things really well: Great developers to crank out stellar code and any sysadmins worth their salt imparting critical availability insights built upon years of operational blood, sweat and tears—shed on everything from big iron to containers. Move people away from their core competency and risk ticking folks off and compromising the business with sub-optimal services.
But with many applications now tightly integrated to their environmental context, more will be expected from every tech discipline. Not least development teams who should understand the performance implications of code running across more fluid and dynamic infrastructure.
It’s here that good monitoring and automation tools can help, but the best solutions go beyond forcing folks to use unfamiliar techniques. In development for example, it’ll be counterproductive to enforce a practice where teams step away from coding to conduct some application performance checking—that’s akin to putting down the fender to pick up a fiddle. It might happen, but if they’re in the middle of a great code riff, what’s the likelihood of success?
To combat this, tools need to become the tech equivalent of musical backing tracks; keeping Dev and Ops playing to the same beat. In Dev, that could involve incorporating application performance management (APM) in context of development activities. So as engineers’ check-in code and conduct application builds, APM can be invoked using their own tools and dashboards.
Melodies and Harmony: Business Metrics and Outcomes
Getting skilled folks working together trumps generalization, but this means ensuring everyone is playing the same tune—business outcomes. However, this isn’t always built into organizational fabric; how many tech leaders present their contributions in business lingo—like revenue, profitability, cost structures, and heaven forbid—the customer experience?
Business metrics are what really counts, so DevOps tools and dashboards should include indicators that enable teams to focus on biggest problem everyone needs to solve—improving the customer experience. This way any one team fully understands the fruit of their labors in real, concrete economic terms and can huddle with others to address both problems and opportunities—some process groupies call this continuous improvement, but if that’s too much of a mouthful, just call it musical jamming.
Sweet Soul Music: DevOps and the Empathetic Organization
Top bands have great musos—be that the iconic front man, the slick bass player or the axe weilding wizard. Individuals yes, but each having an intimate understanding of what makes everything collectively tick. Call it empathy, but great technical teams have similar traits. Rather than blaze off into the development equivalent of a 20-minute unscripted guitar solo, DevOps-centric coders understand how their actions impact others. They’ll not only engineer awesome code, but craft supportability, reliability, testability, secureability (and any number of other ‘ity’ words) into their creations. Again, this doesn’t have to involve slowly picking up new skills, but using integrated toolsets as the glue that bonds the band.
Just as post-punk Joy Division transformed into synthpop New Order and Iggy Pop switched from playing drums to shirtless punkster, sysadmin and developer artistes can always adapt. Yes, that means adding new tech skills to the playlist, but beware of watering down essential talent in the name of DevOps.
In the frantic digital world, longevity demands constant tech revival at an organizational level—not one-hit wonders. Build skilled teams that can always move with the times—which as Bob says—are always a changing.