Wobbly One Of Myriads @wobblewuffess@awoo.space

Fucking exhausted and tired of him making random ass changes outside of working hours. I know he came from a startup... but fuck.

FFff... Senior SRE made some changes to our PD last night around 11pm local for him to alert off of tags...

I got woken up at 3:45 but 4 "Sev 1" alerts paging that were all... in the WARNING category and not supposed to page...

FFFFFF I want to defenestrate our senior SRE right now.

So my coworker wrote a template to handle all the steps of a terraform deploy process (lint, validate, plan... wait for PR before allowing apply) for Azure DevOps Pipelines (which is the tooling we are required to use) so I reused his template, calling it correctly with variables replaced and whatever there are some minor permissions issues I need to fix.

But no... because he can't look at the pipeline.yaml and read it in place to figure out what it is doing (because it calls templates in another repo) he wants me to write a new pipeline from scratch... that doesn't use the templates >.> but also "doesn't want copy pasted work or to duplicate work we've already done"

WHICH fucking one is it?

Sigh... someone needs to tell our senior SRE to NOT FUCKING WORK ON WEEKENDS!

Monday rolled around after our 8am sync up with overseas devs he yanked me in a private call because he completely rewrote our Terraform best practices over the weekend and would like me to rewrite the 4 modules and deployments I spent all of last week working on to conform to his new standards before I go on vacation for a week... in 2 days.

@mawr I saw those things everywhere growing up.. as planters and organizers. I never knew what they were!

Monday Morning.. 3.75 hours of sleep.

Senior SRE has completely replaced Flux2 with ArgoCD on the k8s cluster we were going to hand off to a dev team today in a fit of manic rage last night....

This monday is starting out greaaaat. Not that I mind Argo.. I like it.

IFUCKINGHATEFLUXSOMUCH

I've spent the last two weeks working to get a helm package sorted for a new product dev environment. Today everything was done and it appeared to work

Coworker that built and set up Flux some how configured it to update image tag but not pull the image.

so for almost every way of checking if the changes were being pushed to the k8s cluster and if helm was updating the build image on the surface it was working.

But really it was keeping the same image and changing ALL OF THE FUCKING TAGS TO THE NEW BUILD TAG

I hate this..

Been working on a thing all week and it works, it is down and deployed. Merge all the helm package bits into branch because branch per environment (seems weird to me but whatever)

Senior SRE starts commenting a ton on the Jenkins file... which he wrote and my coworker fucked around with is solely for testing because this is project fork of a thing from that team but we kept the code in the same repo for some reason.
so it has to actually work and shit for PR to be merged.

I don't know shit about jenkins files, I didn't touch it, or write it or do anything with it... but now instead of having a chill friday afternoon of working on some personal development goal training stuff... I'm the one who has to -Fix- this thing because I'm the one who put in the final PR.

So it is clearly my fault... blame and commits aside. But also because both of them are on East Coast time.

And they did not mean had the helm package and chart been tested but they are on the east coast and done for the day... sigh.

If you use some weird word short hand for a process that nobody else knows and it sounds exactly like some other actual process with a tool used.. I hate you.

"Have you done a manual helm deployment to test this?" - comment on a helm release package for Flux2 Helm Controller API

Uh no.. because this is stuff to make Flux understand how to run the Helm API.

The helm package and charts were tested before I started this....

Me wasting a day trying to figure out what the hell they meant....

@Soreth That was similar to my experience but add like serious fatigue around the 24 hour mark that lasted for two days.

Second shot was less bad.. aside from sleeping like 15 hours the day of the shot. Just really heavy fatigue and lots of unplanned naps.

So this morning after presenting a big overhaul that I've been working collaborative with other junior team member on to my old project's engineering team and it going well I get asked to work on Flux thing for new product we are onboarding.

I don't know a lot about flux but it is mostly just yaml bullshitting and I can work backwards from other working examples, especially since bits of it are going to be reused from Other project.

I ask junior coworker if he has any docs for the stuff for this and naming conventions etc he says no.

After a couple of hours I have the basic stuff done for this helm release thing and post up a PR for review. With a comment about how "Since the globals aren't done I don't know the name spaces we are using or service accounts etc, will update them when done."

Other junior coworker has been working on the base globals since last week... complains about how all my values/names/etc are wrong.... then suddenly opens a new branch and commits a huge bunch of shit with about half the base and globals done. With the things I needed... so now I have to go back and re do a bunch of fucking work... and of course the tickets weren't tagged with any sort of depends on...

Like maybe... if that stuff wasn't done.. Senior SRE shouldn't have assigned me the ticket that depended on it? I don't know... fuck.

so far new team is super shitty at ticket scoping. Like at best they put a name in the ticket and a short blurb of what counts as success...

So I'm once again the only west coast member of an east coast team and have to put up with things like "Ideally I'd like to have you around from 10am-4pm EST for collaboration because your quality and velocity of work is very good when you are collaborating with the team."

Me: "I have trouble being functional at the once a week 8am local (PST) meeting to sync up with Euro devs."

Also... like.. you squished me onto a team with the one other survivor from a 6 person SRE team and hired one senior SRE (my team was 4 people + 2 dedicated SWE for our product that we collaborated heavily with) and both of us were the junior members of our teams.. oh and I guess there is the one contractor from the contracting firm but money to keep him ran out so he is gone.

So 3 people + manager... and taking on the work of formerly 12 people with 2 separate managers... and we on-boarding new products to support... by stuffing in the janky other product infrastructure because 2 people from other project out vote me (Manager and other junior SRE)

While my old product and the other product share some technologies (Azure Kubernetes stacks and terraform...) other product is all managed by Flux and had a massive project (mostly done by contractors to update it to the latest Flux)

We passed up on trying Flux as a CI/CD solution because it had a lot of problems with playing nice with the proper way to build Terraform stuff and while that is "fixed" it is still kinda janky.

I'm off on a tangent...

@mawr yeaaah