Another Crack At It
About ~ 8 years ago, I decided to sit down and learn Python. I learned using the well known book/video series ATBS
I know it was ~ eight years ago, because I made this repo eight years ago...
Kazaii- Cringe
- Thanks to the random dude who starred my repo
Reflection
Looking at my repo, it makes me think why I've only dabbled with python over the past eight years. At the time of that script, I was trying to tackle a specific problem. DBIUA had a ISR 1921 router at our edge and only one of our routers had a license for IP-SLA.
Not long after, we migrated to edgerouters; however, I wish I truly believed in other problems to solve at work. The sad part is, most of the network automation I saw was trivial – if not dangerous, in nature.
I've had co-workers bragging about managing ACLs on management switches. Valid & great work, but I wish they focused on draining our routers for all the lifecycles we were doing at that time.
I've seen a very dedicated and talented team deliver services reliably across our CORD modeled fabric. Sadly, I wish they spent more time focusing on testing as application failures could be rather opaque. Especially when their automation caused a widespread outage...
Great people have been doing great things, but nothing truly wow'ed me.
Eureka
Last year I attended NANOG 88 in Seattle and Jeremy Schulman's presentation really captivated me.
I remember working on a project where we were working on network testing for our XR code validation, thanks to all the nasty SR-EVPN bugs we were hitting. Yet, at no point do I remember us considering breaking down the various elements of our production fabric, even when it was duly needed – hell, perhaps we even laughed off the notion.
Now that I think back, our mobility NOC had some great automated testing & tools. The culture was already there and present in our large org, we just didn't know how to tap into it.
Jeremy had vision & drive. I completely missed the moment.
A New Journey
After Jeremy's presentation, I went back to work and started to see the endless downward spiral of our operations. The same issues happen over and over (or at least the same underlying symptoms). Dev & Ops are completely at each other's throats. Issues can be just a few hops away, screaming into the void, but our operators fail to find them.
So I've taken it upon myself to upgrade & catch up my skills in Python, to become a "NetDev" for my org. I'm building a visibility & troubleshooting suite similar to what Jeremy has described in his NANOG88 & Autocon0 presentations.
Unfortunately, I find myself without a mentor & allies to bounce my ideas off of. To remedy this, I'm heading to Autocon 1 in Amsterdam at the end of this month. I'm going to talk python, telemetry, web dev, scalability, IaC, orchestration and much much more. I hope to find my tribe.