Earlier this year I wrote about adding Grafana annotations based on Argo CD events. Since this year I actually got around testing out Flux some more, it was natural to follow up with adding some Grafana annotations based on Flux events.
Clojure multimethods and derivation
I’ve known about Clojure multimethods of course, but I never really used them much. I didn’t really have data that I’d need polymorphism like multimethods to handle, and when I did need something like that I’d use protocols. However protocols dispatch based on class, so when I faced the problem of handling ActivityPub objects that are all maps, it was time for defmulti to save the day.
Structured logging in Clojure
The first time I ran into “structured logging” was way, way back when I first started working on kitsune. Back then I went with the approach of using a Logback Formatter (the JacksonJsonFormatter) to log Clojure maps as JSON. This was nice because I didn’t have to do anything to “transform” the logs further, but had the quite significant downside that logs weren’t really human readable anymore. This plus that that log processors like Loki and CloudWatch can easily parse huge amounts of “traditional” text logs made me pretty much forget about structured logging until recently.
That is until in a thread on the clojurians slack about logging someone mentioned mulog, a structured logging library that caught my attention. I guess maybe from a traditional Clojure background it would’ve been natural to end up with taoensso/timbre at this point, but I never did that bit (though I do have it as an indirect dependency).
The Weave GitOps UI for Flux
Unlike Argo, Flux has no built-in UI. CD tools should be invisible most of the time, just quietly running in the back, getting changes in the cluster. This is fine for “most of the time” when there are no changes to the applications I’m making myself or the cluster itself. It can just look at the various helm charts and other (git) sources and deploying the changes as configured. It’s nice that I can be sure that my cluster has the latest possible versions of everything running within 10-15 minutes of release (depending on the Flux interval configured) without ever touching kubectl.
How do foreign keys work in MariaDB system versioned tables?
When I read about the system versioned (and bitemporal) tables in MariaDB, I was both surprised and delighted. This kind of data versioning is thought of as standard in the “immutable” databases that are big in Clojureland, such as Datomic and XTDB. On the other hand, these databases in Clojureland are usually EAVT based and not the “usual” relational kind. After a little thinking I could think of a bunch of really tricky issues that could arise from system versioning (eg having a by-default immutable log of all changes to some piece of data) in a relational context. The first being foreign keys.
If you create a foreign key in MariaDB without any further specifiers, it assumes ON UPDATE RESTRICT ON DELETE RESTRICT, meaning no referenced key could be changed and no referenced row could be deleted. Just consider this: if there’s a row in some table referencing a row in another, system versioned table, can the row in the system versioned “parent” table be deleted? After all the referenced version of the row will still exist… Or in a similar scenario if the “child” row is system versioned and updated to now reference a different row, will both of those referenced rows be RESTRICTed? After all, both referencing rows will still exist… Of course I had to try.
Hot reloading and reitit in the frontend
Ever since I first ran into the Metosin libraries, I’ve been using many of them. One of the most known and used of those is probably reitit, a routing library that can be used both on the server and in the browser. Recently I’ve had more opportunities using ClojureScript in the browser, and I noticed something that was confusing for a few moments.
Trying FluxCD
When you create a whole FluxCD new setup from zero, it’s really easy: use flux bootstrap. it does “everything” for you. In my case I tried this setup first for last year’s advent calendar. Except back then I had an expedition scheduled for the first half of December, so this took the back seat and was eventually forgotten. Therefore in this year past everything I did back then slowly sank into oblivion, so I again had to start pretty much from zero.
I did set up the repo and a cluster (on Civo) back then, but I quickly tore down the cluster when I realized I wouldn’t be able to test it all out. The repo stayed though, so now i was starting afresh with a k3s cluster spanning three nanodes (that’s a linode referral link with a $100 credit over 60 days) and a repository already set up from last year.
Prometheus vs Longhorn, fight!
Last year I had a brief affair with Longhorn. It’s a tool that abstracts the interface to interact with volumes (in my case in Kubernetes) from where the underlying data actually lives. In my case, my cluster consists of three small nodes, and back then most of the data lived in local-path provisioned volumes. Using local-path means that the data is physically on the host machine instead of some virtually mounted filesystem. This also means that once an application has a PVC, it can’t be assigned to any other node (it results in a “conflict”).
Fuck ads
There are ads that I don’t mind. I’ve multiple times clicked on ads on Facebook and ended up actually buying stuff I’m happily using to this day (trousers, shoes, games…). On the other hand I don’t remember ever clicking on any Google ads, mostly because I adblock everywhere by default so I don’t even see them. On my work machine, where I don’t have adblock, Google ads are just plain stupid so it never even occurs to me.
Maybe a new watch
I’ve been using a Garmin Instinct since I got back from Aconcagua, both for training in the gym and running or hiking outdoors. I might need to replace it sooner than later.
