shahinghasemi

But if you want to be valued you must be useful at the first place.

- zoom in your screen/font. - use blue-cut glasses. - increase your distance with the monitor. - break from monitor when your eyes are tired. each 20 minutes spare into an object far distance. - drink water.

As you mentioned there are easy workaround for this issue on both application and infrastructure level: implementing rate limiters, monitoring/alerting systems for DB resource usages, putting quotas for tenants, etc. After all a postgreSQL instance in 16GB ram and 8core cpu can handle pretty much of the traffic for most SaaS products at their first stage.

What misbehaving do you mean? If traffic is what you mean, it's okay since at the first stage of the software we have time to test/scale accordingly. Therefore in that stage it might make sense to change the architecture to per-tenant database.

It differs for each tenants. Preferably there should be no limit for the number of orgs/tenants. The traffic for each tenant is also varying but it's safe to consider 500req/s.

The main reason: getting rid of synchronization/migration process for many databases.

What fish oil are you using?

Without discipline you get to nowhere. At the start discipline is hard because we are a human which our brain wants to do the easiest task available to preserve energy. to prevent this situation we have to train our brain to obey us not otherwise. repeating this pattern lets your mind knows that you are the boss not itself.

The critical part here is *you have to separate yourself from your mind/brain*, in other words you are not your mind/emotions. You can control them or otherwise being over controlled by them.

Have not read the article yet but I guess it can be solved using incremental tests. First we write the tests for the core features of the API then while implementing it to pass the tests, we'll find that there arw some edge cases that we have not consider, so we write tests for them. In addition tests framework like jest provides good API around the coverages.

To compare them, would you please elaborate on the technical side regarding the write throughput?

This is my learning strategy which originally posted on [1]:

I have created a learning framework for myself to learn literally anything. First of all It’s not going to be easy to get started learning something new because there are many holes in your mental model about the material and the aim of the learning is to fill those holes. We’re living in an information era. You can drawn yourself in surfing the web/docs/articles if you are some kind of perfectionism to convince yourself that you’re learning. But that’s the trap. You should spend about 30-40% of your time learning the material and 60-70% of your time applying it and then under the process you’ll find: “Oh I don’t know how or what is X” then you go to learn about X specifically. This works best. In other words as you will be applying what you have learned so far you start “learning on demand”. In other words filling the individual holes. This loop will force you to learn the topic.

[1] https://news.ycombinator.com/item?id=29116026

At NodeBB, we ended up relying on websockets for almost everything, which was a mistake.

Would you please elaborate on the challenges/disadvantages you've encountered in comparison to REST/HTTP?

I have created a learning framework for myself to learn literally anything. First of all It’s not going to be easy to get started learning something new because there are many holes in your mental model about the material and the aim of the learning is to fill those holes. We’re living in an information era. You can drawn yourself in surfing the web/docs/articles if you are some kind of perfectionism to convince yourself that you’re learning. But that’s the trap. You should spend about 30-40% of your time learning the material and 60-70% of your time applying it and then under the process you’ll find: “Oh I don’t know how or what is X” then you go to learn about X specifically. This works best. In other words as you will be applying what you have learned so far you start “learning on demand”. In other words filling the individual holes. This loop will force you to learn the topic.

It's a Persian idiom and what does that mean is that you can't hide everything from others and they'll be informed in such way that you'll be surprised(don't know how!).