Inspect SafeEthClient for concurrency issues

[emlautarom1]

We already found a couple of issues related to the SafeEthClient (see #172, #173) but there is a high chance that there are more lurking around.

We would like to revisit the code in the following package: https://github.com/NethermindEth/near-sffl/blob/8d1dc75c45cdfc9ea26169f2bd1816a597582e1f/core/safeclient/client.go

[Hyodar]

I think this is still relevant. Still, since this is probably the best place to keep this discussion, I feel like my concern on concurrency may not have been well-founded in that case. Basically, the context is these downtime issues happened close to the operators updates in kind of a random way. At this point, we're kind of in the mercy of the RPCs, so not being sure about that is problematic. Still, for quite some time already things have been fine - only a short downtime in one operator and an operator that did stop tracking blocks at a point is actually running the previous version, which did exactly that (no re-sub). The relayer, which is now using the safe client as well, has also been running fine.

[Hyodar]

My proposal for now is to keep an eye on the metrics (both #176 and #177 are improvements to that) and also merge #175 so we have something easier to work with. The previous reinitialization design makes things a lot harder than they need to be in an unnecessary way - what we really only needed there was resubscribing from time to time and that's it.