Slow/unresponsive HA fixed by ha network reload - but why?

This is sort of a weird topic. I am not asking for help, because I seem to have fixed my issue, but I’m not sure why the action I took fixed it, and I’d like to understand.

Been running HAOS (Raspi3, 64GB SDHC class 10 card) for about a month now. Vanilla HA installation, except for SmartIR, which I installed myself by cloning it into my custom_components directory.

A few days back, HA started to be very slow to respond generally, but especially any dashboard that had a state history on it. Seemed to take forever to pull up history data. Situation got worse and worse, until finally, it seemed that I couldn’t connect at all. I did a reboot or two, and that seemed to help a bit, but even as the HA instance came up, it was SLOOOWWW and highly intermittent. This is evident in my history graphs - lots of gaps in my plots, and just general unpleasantness trying to connect.

I checked the logs for core, supervisor, etc. Nothing in there seemed especially damning, except for these “XYZ took more than 10 seconds to execute” which seemed like a symptom rather than a cause.

Googling an error I found in the supervisor log, which indicated that the supervisor had no internet (this eemed like nonsense to me, as i could access the instance through HA cloud, just the same as local) - I did a

ha network reload

which was indicated by a forum user somewhere.

This seemed magically to fix my issue. My instance is once again nice and snappy, and has been that way for hours now. I have done a physical reboot, and the fix seems to have held.

My question is: What gives? What did the reload actually do - why did that fix my issue? I’m just hoping to learn a little bit here about what’s going on for next time. Like I say, my issue seems fixed, but i don’t feel like I’ve grown

haha, well, maybe I spoke too soon. After a few hours of good performance, it’s back to being highly intermittent and slow. Back to the ol’ log mines, I guess!