The 21 Day SPAUTO

I'm dumb. Let's just get that out of the way now. No need to @ me.

You would think I was dared to do this. Or challenged in some way. But no. For some silly reason I got it in my head that I would try and do the SPAUTO 300-535 exam in three weeks. Start to finish.

And the TL;DR is that today I passed the exam. But if you wnat to know how, then keep reading.

Just Minding My Own Business

You see, I've been thinking a lot lately about going after the DevNet Professional cert. I've actually wanted to for a very long time, but for $reasons I've never seriously attempted it. I have, however, done a few things that align with that goal such as learning to be competnet with Python, work with a few APIs here and there, and looked at many automation technolgies in at least a limited capacity. So I'm not starting from nothing to be totally honest, but before January 12 2026 I had never cracked open the Exam Topics for SPAUTO, or ever actually considered that might be something I would attempt.

So what changed?

I've had a renewed interest in learning recently, and the aforementioned Devnet Professional was the decided goal. However, DevNet Professional is (was by the time you read this) a dead person walking. As of February 2, 2026 DevNet Professional is no more, to be replaced by the all new CCNP Automation. New name, and more importantly, new exams. DEVCOR was being retired, along with all the "auto" specialist tracks save for Enterprise and Datacentre.

Honestly, this was kind of a bummer for me. I had previously thought I would go after the DevOps 390-910 specialist track but was being retired so I was out of luck on that. I had resinged to the fact that I would likley being going after the new ENAUTO 300-435 v2.0 to complete my CCNP Automation after I first completed the all new AUTOCOR.

But I found myself in a bit of a lull period. It was early January and I had to wait till February for the new content on Cisco U to drop. Yes, the AUTOCOR exam topics had been released, but I was also mindful that my existing CCNP (Enterprise I think It's called now?) was set to expire in June. I had a few CEs already accumulated and I was hoping to get the remaining that I needed for my recert. So in my mind the clock was ticking. I felt that sitting around for three weeks wasn't ideal.

So there I was, looking at the Cisco U catalog, trying to figure out my next move for CE credits. That's when a voice said to me "How about the SPAUTO?". The what now? Why the heck would I do that???

This was on Jan 12. SPAUTO (and everything DevNet Professional for that matter) was all set to be retired Feb 2. Literally 3 weeks to the day.

I checked the Exam Topics. They didn't seem that bad. Thinking ahead, after CCNP Automation the only other Cisco track that even remotely interests me is the Service Provider track. SPAUTO could actually be a good move for me to set myself up for both certs this year. I chatted a bit more with that voice and gave it some thought. Hmmmmm.....

OK, screw it. Let's do this.

I got enrolled in the SPAUI Cisco U course (24 CEs) and booked the exam for Feb 1., one day before it was set to retire. One chance to do this. No turning back.

When The Rubber Meets the Road

Thankfully I'm set up to take on a task like this. As mentioned, I was able to get the Cisco U SPAUI course set up right away. I also have my own CML instance at home to build and test with (I'm looking at IOSXR here mostly), and another "server" that I use for KVM/Docker that's ready to go when needed. That server just happened to have VM dedicated to NSO already set up and available. This was a great starting point and saved a lot of time that would normally be needed to spin up a lab for something like this.

What I didn't have that I did need to spend time building was:

• Ansible control node
• Telegraf/InfluxDB/Grafana (TIG) stack

Which, really wasn't much. The TIG stack easily took the most time, but seeing as hwo that's the only real heavy lift here I was looking good. The TIG stack took me 2-3 horus to get installed and running, but the Telegraf config to actually handle telemetry stream took much more time. I probaably spent the most time on that one item, and no part of the Exam Topics are asking you to set up Telegraf.

Ansible did also take some time to get an inventory set up, and some playbooks (I'm an ansible n00b I'm ashamed to say) but since those are Exam Topics that time was much better spent.

Alright, so that's the homelab side. What about the content? Surely the Cisco U course wasn't sufficient?

No, it was not sufficient. Once I finished with the SPAUI content I went looking for other content.

I immediately found https://github.com/h4ndzdatm0ld/spauto_devnet. This looked promising, but in the end I didn't use this resource too much. I already had my own lab, and the notes were fairly incomplete. But something was definitely better than nothing.

Digging further I found some references to Nick Russo's old SPATUO study plan, but his website unfortunately no longer exists. But with a little more searching it turns out a lot of his stuff is on GitHub. Here's teh SPAUTO directly:

https://github.com/kaelemc/NickRussoContent/blob/master/spauto_studyplan.xlsx

I'll just pause here and say that even though I never met Nick, I followed a lot of what he did and he was someone that I admired tremendously. So many people have benefitted from the content Nick produced, and I'm very happy to count myself in that crowd. His SPAUTO study plan may never get used again now that the test is one day away from being retired, but I can truthfully say that it did get used right up until the day before the exam was retired.

I don't have Pluralsight access, so I wasn't able to access Nick's own content in that study plan, but there's a ton of other links to reading and videos that are very relevant to the SPAUTO. I read/watched as much as I could.

What else? Not too much really in the way of learning content. From here is was all about fingers on the keyboard and making sure I knew how things worked. So with an open terminal and a bash shell it was time to hit:

• netmiko
• ncclient
• NCS MAAPI and MAAGIC
• pyang
• ansible

I never actually set up YDK, but I felt like I hit up enough resources to know what the code would look like and what all the components were. I couldn't bring myself to acutally install and use it.

On the IOS side, I already has IOSXE devices running in CML for NETCONF/RESTCONF stuff, but I didn't have IOSXR devices running. That was easy to fix and then I had those for NETCONF/RESTCONF as well. I also ended up using IOSXR for some GRPC and gNMI learning as well.

The last part that comes to mind right now (i'm probably missing something) is SR-PCE (Formerly XTC). I made sure I was good with getting a PCE set up and talking to PCCs with PCEP. I also made sure I was good with BGP-LS.

Side note, I didn't know BGP-LS was a thing and for whatever reason I though that was actually pretty cool.

To wrap the prep stuff up, I'll say that I did do some light reading of the relevant RFCs but nothing too serious. Mostly just some skimming of them to get a feel for what content was there without actually killing myselkf over drilling the actual content into my head.

Also, Shout Out to an old colleague and running buddy of mine, Mr Brian Yaklin. When I was messing around with ncclient I got stuck with actually trying to save the config on an IOSXE device without a candidate datastore enabled. Googling led me to his blog where I found the answer.

https://www.yaklin.ca/2021/03/18/exploring-netconf-and-yang-on-cisco-ios-xe/

Am I Ready? It Doesn't Matter.

Since the exam is now retired I'll take a few liberties with my impressions.

Before taking this exam I had wondered how Cisco would test your code knowledge. In other Cisco exams you encounter Simlets and Sims that put you into a simulated environemnt and ask you to configure a specific technology/scenario. It's probably just my Google-Fu failing me, but I couldn't find anything great about what to expect. Just this Cisco blog with what seemed like unhelpful text:

https://blogs.cisco.com/learning/new-performance-based-lab-exam-items-build-opportunities

Yeh, to general, and not really applicable. Would have been nice to have something a little more direct for this exam.

As for the exam itself, I'm assuming that if I'd taken other DevNet exams before then I would have been better prepared for the way Cisco goes about testing coding knowledge. I was definitely suprised with exactly how much of the exam tested knowledge of code syntax. The SPAUTO exam topics lists items such as ncclinet, netmiko, and YDK. If you don't know what code using those Python modules looks like, you're cooked.

The same can also be said about using requests to interact with RESTCONF. Knowledge of header and body construction, and the syntax of the various http methods was an absolute asset.

The other heavy hitter on the exam, which I think was to be expected, was YANG. You sure did need to have a handle on YANG! I will freely admit that I think I fell short in this area. Sure, I'm comfortable with the basics, but this exam went further than that. I burned the most time on the YANG questions that tested deeper knowledge, and it's those YANG questions that left me scrambing at the end to complete the exam. The reason I spent so much time on the YANG questions was tryingt to reason out the answers using the incomplete knowledge that I posses. Since I passed I can only assume that I was able to guess reasonably well. The takeaway here though is that I definitely need more time with YANG.

Looking at the exam topics in general, the exam did hit nearly everything in some way. Based on what I saw in the exam the exam topics was accurate in laying out Cisco's expectations of you.

Final Thoughts

So, all that is to say that I actually liked this exam, with a few caveats.

First and foremost, the biggest drawback to this little exercise is that I willingly --spent-- wasted time learning YDK and YANG Explorer knowing that these tools are completely dead. This is definitely regrettable, but to keep things in perspective I've also spent countless hours learning technologies liek X.25 and Frame Relay for Cisco exams. Those are also long dead now. The difference being that when I learned them they were still in use, but at the end of the day the result is the same.

It's still a Cisco exam that carries with it a certain amount of that Cisco exam bullshit. All in all though I feel like this exam was light on that and the questions were worded well with reasonable choices. I absolutely hate the "best choice" type of question where all answers are basically wrong, and you have to pick the most right (whatever the hell that means).

Overall I did very much enjoy the content of this exam and I'm glad I was able to get this one under my belt.

AUTOCOR, you're next!