Rupesh

Happiness of decreasing the response time 10 times which gonna impact millions of users is unbelievable 🚀😎🚀 Thanks @newrelic for such an awesome monitoring tool.

@ZeptoNow @jagograhakjago any update on this ?

@ZeptoNow Order got cancelled by delivery agent and money is not refunded yet.. its been 2 weeks Shitty exp has no customer support on app,shitty chatbot.. If you dont take the action on this.. I am going to report in consumer court. Order id = 83655BDNM10867

@__alter123 @Tweetbyjagan @SumitM_X You dont need to worry about data loss and sharing data over network, and how big is the file.

@Tweetbyjagan @SumitM_X what makes S3 more efficient compared to direct transfer?

Your microservice needs to transfer large amounts of data (e.g., files, images) between services. How do you design the communication to avoid performance bottlenecks and manage large payloads efficiently?

@Tweetbyjagan @SumitM_X Ideal one

There must be chad level practice as well behind this.. op confidence on skills.. 🫡🫡

75% of developers will fail this SQL question. What's wrong with this script? The first time I faced this question, I jumped into the SQL syntax like crazy. But the syntax was OK; the problem was in the data and related to a concept called Cardinality. Cardinality refers to the number of unique values in a column relative to a table's total number of rows. • High Cardinality means the column has many unique values. • Low Cardinality means the column has few unique values. Creating an index on a column with low Cardinality is most of the time ineffective because: 1. Low Cardinality means each indexed value points to many rows, reducing the index's ability to narrow down the search. 2. Maintaining an index has a cost of storage and update time. For low cardinality columns, this overhead might outweigh the benefits. 3. Database query optimizers are smart; they know column statistics, including Cardinality. When they detect a low cardinality index, they often ignore it and perform a full table scan instead. A simple Example Consider a table "Employees" with 1 million records. Let's examine indexing on different columns: • ID: High Cardinality (1 million unique values). An index here would be very effective. • Name: High Cardinality (many unique names). An index could be helpful in searches. • Department: Medium Cardinality (10-20 unique values). An index might sometimes be useful but less effective than an ID or Name. • Gender: Very low cardinality (2-3 unique values). The query optimizer would likely ignore an index here. When to consider a Low Cardinality column? There are scenarios where indexing a low cardinality column might be beneficial. For example, combining low and high cardinality columns can be effective. 'CREATE INDEX idx_dept_emp ON Employees(Department, ID);' Now you know the basics about Cardinality! Save this post; it might help in your next interview.

@pankaj013 @Franc0Fernand0 "I have failed more times than you have tried"; So I can deduct the failure patterns quickly.

If you are a junior engineer, you may need to hear this. Senior engineers aren't smarter or faster than you. They've simply seen the thing you have been struggling with many times before.

I am thrilled to announce that @ManningBooks has decided to publish Thinking in Distributed Systems Writing this book was challenging but rewarding. Your support, your feedback and comments kept me motivated throughout. Thank you ❤️

These Lines 💯

@the_dream_saver Seems too much contribution in small cap

Who said mutual funds don't make money?

No love in this world is bigger than a mother's love

@ProgressiveCod2 Have one doubt - what if the concurrent code is running on multi-core processor, then that will be treated as parallelism right ? Because cpu have the capability to handle the things seperately..

Concurrency & Parallelism are two terms that always create some confusion. Here's how I try to understand them: 👉 Concurrency means more than one task can appear to make progress over a unit of time. The key term over here is "APPEAR". Concurrency gives the illusion that the tasks are running at the same time. In reality, the processor is switching between different tasks. 👉 Parallelism is when two or more tasks actually make progress at the same time. The key word is "ACTUALLY" For parallelism, you need to have more than one processor or a multi-core processor. Multiple tasks can make progress at the same exact time. But that's not all. In reality, tasks can be executed in several ways: - Sequential - Concurrent - Parallel - Parallel & Concurrent 👉 A few important points to remember: [1] Concurrency is about dealing with lots of things at once. [2] Parallelism is actually doing lots of things at once [3] The application can be concurrent but not parallel. [4] The application can also be parallel but not concurrent [5] All parallel executions need not be concurrent 👉 So - what are your views about parallelism and concurrency? Do you think there are more differences between the two?

How GitHub manages their databases in production? ⚡ Database outages are more common than we think, so what happens when the master goes down? how a replica is chosen and promoted to be the new master? Is this automated or done manually? GitHub published a blog talking about the tooling they use to get an optimal DB performance and High Availability and I dissected it to and compiled my learnings in a video. give it a watch - youtu.be/4mVJQJbw6Vw ⚡ I keep writing and sharing my practical experience and learnings every day, so if you resonate then follow along. I keep it no fluff. youtube.com/c/ArpitBhayani #AsliEngineering #SystemDesign #DatabaseEngineering

@arpit_bhayani Had experience exactly same issue in past but the root cause was something diff.. how we handled is took downtime of 5 mins and closed the requests coming on db from gateway for that interval and then manually up the master.. and then enabled gateway.. and it worked likr charm 😅

@arpit_bhayani @kritikatwtss Agreed!! To scale write you need partiotining and sharding depending on the use case, @arpit_bhayani has explained this in one of his video as well, you can refer that!

this is incorrect. master will "only" accept write is wrong. writes and critical reads will go to the master. the reads that need a consistent view of the data need to go to the master. there is no way out. also, to scale writes beyond a single machine you need to partition the data and distribute the ownership across.

From Alex's system design interview book: Database replication: use master-slave architecture master will only accept write operations, whereas the slave will have read operations (read-only). Since most of the applications are read-heavy we have more slave nodes than master nodes. But what if the write operation is heavy for the application? How do we take care of this?

@Priyanka__488 Can you share the blog link ?

who could be more excited about garbage collection ✨