Rohit Choudhary

Which all metrics needs to be monitored for restful webservices? 1. Response Time: Measure the time taken to process incoming requests and generate responses. Monitor average response time, percentiles (e.g., 90th, 95th, 99th), and trends over time to identify performance issues and bottlenecks. 2. Throughput: Monitor the rate of incoming requests and processed responses. Measure request and response rates per second or minute to understand traffic patterns, peak loads, and capacity utilization. 3. Error Rate: Track the percentage of requests that result in errors or failures. Monitor HTTP status codes (e.g., 4xx, 5xx) and error responses to identify issues such as client errors, server errors, or network issues. 4. Availability: Monitor the uptime and availability of the web service. Track service availability over time and calculate uptime percentage to ensure that the service meets reliability requirements and SLAs. 5. Latency: Measure network latency between clients and the web service endpoints. Monitor round-trip latency, DNS lookup time, TCP connection time, and time to first byte (TTFB) to identify network-related issues and optimize performance. 6. Concurrency: Monitor the level of concurrency or concurrent connections to the web service. Track the number of active connections, connection pool usage, and queue length to ensure that the service can handle concurrent requests efficiently. 7. Cache Hit Ratio: If caching is implemented, monitor the cache hit ratio to assess the effectiveness of caching strategies. Track the percentage of requests served from the cache versus the origin server to optimize cache performance and reduce response times. 8. API Usage Metrics: Monitor usage patterns and trends of API endpoints. Track the frequency of API calls, resource usage (e.g., CPU, memory), and API usage quotas to understand usage patterns and plan for capacity scaling. 9. Security Metrics: Monitor security-related metrics such as authentication failures, unauthorized access attempts, and anomalous behavior. Track security events, audit logs, and compliance metrics to ensure data protection and regulatory compliance. 10. Dependencies: Monitor dependencies and third-party integrations used by the web service. Track availability, latency, and error rates of external services, databases, and APIs to identify dependencies that may impact service reliability or performance. CONCLUSION:- By monitoring above mentioned metrics, issues can be detected, performance can be optimized, and ensure the reliability and availability of their RESTful web services. Some common tools available for monitoring these restful api are available like Prometheus , Grafana and many other. #backendEnineering #apis #systemdesign

6

🚩 Red flags in a Software Engineering opening ❗️ Listing EPF as benefit (It's mandatory for all employers with staff strength > 20, its a govt rule and not a favour.) ❗️ Listing medical insurance as benefit (It's mandatory for all employers in India, it's a govt rule and not a favour). ❗️Asking for expertise in 20 different tech stacks (I believe nobody is looking for a marvel character) ❗️Explicit mention of "competitive" salary and then asking for salary expectation. (can we blame candidates for negotiating other offers?) ❗️Employee Assistance Program (am I the only one to hear this only at the time of layoffs ? ) ❗️Labelling the job as "entry level" and then write 2+ years of experience in the JD. (entry level and junior levels are entirely different) ❗️Tier 1 labelling (I wish one day they also come up with a logic saying all candidates named "Rahul" can apply, because we believe they're competitive and smart 😎) #jobs #career #hiring #recruiting #engineering #softwareengineering

79

3 Comments

#Paytm employees have taken to the Ministry of Labour and Employment after their unfair layoff without any proper severance package. We all have been reading the brutal layoffs news for the past 2 years; one where the CEO just shows up on the Zoom call to fire 80% of the company or where security guards show up after a conference and employees are escorted. In some cases, they find a legal loophole by forcing you to admit that there was a performance issue and putting you in PIP and thereby crushing your self-confidence and then saving some money by not paying their dues after firing you. There are companies that treat their employees like trash and expect the general public to buy the narrative of "We are a family". If an employee absconds a company then it is frowned upon in the industry, the companies scare you with a bad relieving letter or failed background checks in the next company and yet these same companies face no consequences when they happily ruin the financial and mental status of their so-called "family members". With privatisation on the rise it is high time that we all voice for better labour laws in #India so that people can work peacefully. Because we work to live and not live to work. #worklifebalance #layoffs #labourlaws

22

Below is a list of companies that offer a base salary of 10L+ to SDE 1 and most of them offer 20L+ base salary to SDE 2, as per Glassdoor, LeetCode, and AmbitionBox. I can definitely vouch for the companies : #PayPal and #Tesco here, but do your due diligence for other companies when it comes to work-life balance, work culture etc. P.S: List has backlink to company page, which is clickable only when the pdf is downloaded. #jobs #hiring #softwarengineer #sde #engineer #salary #ctc #linkedin #career #openings

365

8 Comments

After watching Akshay Saini 🚀 's "Will AI Replace Software Engineers ?" video I agree with the insights shared, which serve as both a learning experience and a wake-up call. How will engineers loose their jobs in the future:🛩👎 1. Over-reliance on ChatGPT for tasks. 2. Skipping debugging processes in favor of immediate AI or internet-based error resolutions without critical thinking. 3.Being constantly in your comfort zone without seeking challenges. 4. Neglecting upskilling in current frameworks or emerging technologies. 5. Lacking curiosity and enthusiasm for self-improvement. 6. Having a negative,give-up mindset instead of actively working on self growth 7. Disregarding constructive feedback. How to survive this AI wave successfully:✈️👍 1.Don't rely extensively on ChatGPT for every little thing. You might be thinking I'll use it today what loss can it have but few years down the lane you will regret and it would have become a habit by then. 2. Maintain a strong foundation in your core technical stack. 3. Continuously upskill and explore new technologies. 4. Strive for excellence, recognizing the need to surpass average standards in the face of AI advancements. 5. Understand that while job opportunities may decrease, exceptional developers will survive. 6. Cultivate a positive outlook and adaptability. 7. Being an expert and having in depth knowledge in 1 framework/language is way better than being average on 10. 8. Emphasize problem-solving skills over just coding. 9. Write and debug code on your own to actually learn and improve yourself. 10. Ensure that AI compliments you at the same time you not being overly dependent on it. Cheers!! . . . #software#softwareengineers#AI#articles#akshaysaini#namastejavascript#javascript#angular#react#framework#html#css#frontend#developers#designers#ux#figma#growth#networking#tuesdaytalkies

16

Bangalore? Gurgaon? No thanks. Remote job + travel is one of the best things you can do to your life as a developer. - Less stress, more productivity - Less chaos of big cities, more spending power - No time wasted in travelling through a hoard of traffic every day #remotejobs #motivation #developer #programming #remote

104

14 Comments

I am really excited to share with you all that I have started a new series on Spring Boot Tutorials from Scratch. In the first video I have covered the basics of Spring Boot, including its key features and benefits. Video Link: https://lnkd.in/gXQFPstU We also have compared it to traditional Spring to see how it streamlines the development process. Don't forget to like, share, and subscribe for more coding tutorials! #Repost if you find this video useful Topics Covered: What is Spring Boot? Simplified Setup Stand-alone Applications Production-ready Features Rapid Development Auto Configuration Spring Boot Starters Spring VS Spring Boot Setup & Configuration Dependency Management Embedded Servers #lazyprogrammer #spring #springboot #backend #developer #java #youtube #like #share #subscribe #repost

4

Just going through the NamastheyJS series from Akshay Saini 🚀! Found out that 'Hoisting' is not an explicit term defined in ECMAScript! It's more of considered as a default behaviour of JS rather than some special feature. More on this: https://lnkd.in/gbW7aMFJ #JS #frontend

14

2 Comments

Stripe processed more than $1 trillion payment transaction volume with an availability of 99.999 %. DocDB, it's underlying database plays a pivotal role in Stripe's high scalability and availability. 🚀 🚀 Let's unlock the secrets of DocDB's high availability and scalability & dive deep into its architecture in this post. 📚 📚 What is DocDB ? 🤔 🤔 DocDB is an in-house database-as-a-service (DBaaS) built by Stripe. It internally uses MongoDB as the database engine. Stripe teams rely on DocDB's APIs to store, and retrieve data. What were the key requirements that led to DocDB ? 👉 High availability, performance and durability 👉 Horizontal scaling with sharding 👉 In-built support for multi-tenancy with quotas 👉 Strong security through authorization Why did DocDB choose MongoDB as database engine? 💡💡 ✅ Document-model offers flexibility. ✅ Handles large throughput of real-time data at scale. ✅ Better developer productivity over SQL databases. What are the key components in DocDB ? 1️⃣ Physical database shards - These consist of compute and storage nodes where the data gets stored. 2️⃣ Database proxy - This layer handles the client queries, and routes it to one or more shards. 3️⃣ Chunk Metadata Service - It handles the mapping between the chunk and the corresponding physical shard. 4️⃣ Change Data Capture (CDC) pipeline - It captures all database modifications, converts it into events and stores it in an object store like AWS S3. How does DocDB handle a client request ? 🎯 🎯 - Client uses DocDB API and sends a query. - Database proxy parses and validates the query. It sends a request to the Chunk Metadata Service to lookup the physical shard corresponding to the data. - Chunk Metadata Service returns a list of one or more shards to the DB proxy. - DB proxy then sends multiple parallel requests to different physical shards. - DB proxy waits for the response, aggregates results from multiple physical shards and returns it to the client. What are the striking features of DocDB ? 🔥 🔥 🌐 Handles 10K+ distinct queries to deliver petabytes of data. 🌐 Supports 5K+ distributed collections. 🌐 Distributed data across 2K+ database shards. 🌐 Provides 99.999% availability. DocDB uses a famous pattern known as scatter-gather to fetch the data. Let me know in the comments below if you have used this pattern while building real-world applications. 👇 👇 In my next post, I'll dive deep into the DocDB's data movement platform that enables data movement without any downtime. 🔨 🔨 You can find the link to DocDB's blog post in the first comment. 👇 👇 For more such posts in tech, follow me. #tech #databases #SystemDesign #api

73

2 Comments

As a senior software engineer, ensuring a strong grasp of system design is crucial for excelling in technical roles. Here are some key points to keep in mind when preparing for your next system design interview:📌 Architect scalable solutions with confidence.📌 Optimize performance like a pro.📌 Understand problems with clarity and design optimal solutions. Embrace the challenge and soar to new heights of success! Follow SIMRANPREET KAUR for more insightful content on system design and tech leadership. #SystemDesign #SeniorEngineer #TechLeadership

1

Whenever you see any post like - "How i got into sde1/sde2, Are you worried about DSA & SystemDesign" 99.99% It must be a promotion of either Postplaced or Losscoder. Nowadays most of the engineers be like - Join FAANG and earn more money in promotion than your job. The reality is - Most of them didn't take any paid course to achieve that. Some people will open youtube channel immediately and will post some useless videos. Like - " How a google software engineer spent salary in Bangalore? " Note - There are some genuine channels also which helps really well. #stoppromotion #savemoney #Learnandgrow #sde1 #sde2 #dsa #systemdesign

9

My Linkedin is the cleanest it has been since 2020. No sponsored post, no TCS to FAANG story with link , no DSA and stuff. Even if someone tries to tag me in someone's post they have already blocked me. Yesterday someone DM'ed me "Look at this brainless influencer promoting this product and what not" . I clicked on the link just out of curiosity and nothing 😂 . You can say I am now using Linkedin premium with ad-blocking enabled. And I acheived that with just 5 posts 😂 😂 . If you are tired of these posts highly recommend this instead of going over and manually blocking hundreds of influencers. Better try being in their block list. PS : Aur yaar padhai karni hai to youtube pe aa jao. Yesterday we did a system design of whatsapp and hotstar : https://lnkd.in/guSFJkCP Finally I know some people are walking my path unko bolunga yaar negativity mat failao , if you feel an edtech is scam try building a better one. The only reason students fall for scams is because they don't have better option. Most students choose to pay 1 lakh for DSA as a last option kyunki unhe dhang ke mentors nhi mile.

31

1 Comment

Hi Amazon team, U guys sucks... a simple subscription model releases with P0 and P1 issues... u guys coded enough so that anyone get easily pissed off from renewing prime video account.. Is subscription team built by Interns ? try out: prerequisite : Accounts prime membership has expired ( purchased last year) steps to reproduce: 1. sleect a movie which are offered through Prime ( over Firestick) 2. click Watch now.. and a a nice QR code for purchasing membership will appear but thats a "Catch me if you can " game .. where u will see the different tier of memberships.. and thats all My 2 cents ========= 1. When I m coming to purchase membership from Prime Video flow.. why u r showing me ur shitty Prime cloth wala subscription? What was the idea behind this brilliant design , sorry I didnt get that.. I m a general user re... seating with my family to spend some light moment and u have perfectly designed and coded to ruinthat out. 2. If the design was as per Leadership / Buisness ask to keep the MVP rooled out at earliest then take my Salute to great those leaders.. may God bless you , and sometime pls think beyond your business tagets and Test it well atleast P0s.

1

Look, at how a company's quarterly performance may impact a common person's life 📉 Paytm, a big player in financial technology (fintech), faced challenges in the last quarter of its fiscal year 2024. 🔴 Operating revenue dropped to 2267 crore in March 2024 from a high of 2850 crore in December 2023. 🔴Contribution profit also went down by 15% QoQ, reaching 1288 crore in March 2024 from 1520 crore in December 2023. 🔴 EBITDA (Earnings Before Interest, Taxes, Depreciation, and Amortization) before ESOPs decreased significantly from 219 crore in December 2023 to 103 crore in March 2024. PayTM experienced a challenging period with declining revenue and profits in the last quarter of FY24. These financial indicators show a significant decrease in performance, reflecting the company's current struggles.  Result is Layoff!!!!

17

3 Comments

You can succeed in your career even if you messed it up. I published a latest edition of my newsletter which showcase how I still got promoted despite bringing down Amazon Prime 1-day delivery promise. Yes, it was my bad code. Here are the key takeaways: 👉 Build trust by being a go-getter and meeting leadership expectations, and they'll support you in tough times. 🙌 Always learn from your mistakes, take the lead, set examples, and come out stronger. 👌Build trust with business leaders, keep them informed, and help them achieve their goals. ❤️ Strong leadership is a blessing. If you have supportive leaders, maintain lifelong relationships and strive to become a similar leader. If you want to read the full article, the link is in the comments.

5

1 Comment

“Bhaiya at the end to react native ka code bhi jvm bytecode m hi compile hota h na?” This was a response i got from a developer when asked about how react native works in Android. But reality could not be farther. See, Android DVM is not capabale of running javascript code because these are inherently different technologies. Instead, what happenes is that a JSC (JavascriptCore) runtime that is bundled into android application alongside react native libs which actually loads the javascript code into memory at runtime. And then, there is a React Native bridge native library (.so file) which acts as a intermediator, translating the calls from java/kotlin code and js code. And because of all these intermediary layers translating instructions among languages, the performance for react native is not as good as native applications. Hope it helps.

22

Sometimes I am amazed by the power we have as Software Engineers. Recently due to deploying the wrong configuration change in our service, we had a high impacting issue due to which all Fulfillment centers operations in NA and EU across Amazon shut down for almost 20 mins. Below are the learnings i had from this incident: 1. Any configurations being deployed manually will always be prone to human error. 2. Setting up processes and SOP will only take you so far. Well SOPs are good for short term but in long term there should be an automation on how these configurations can be deployed. 3. If your service is being used across geographies ensure you have separate stack for each of the broader regions. The traffic pattern will be different due to difference in time zones. Peak traffic at one region can be a downtime for other. Ensure you deploy your changes in downtime traffic of the respective regions. 4. If you are making any change for a particular use case don't assume existing use case will not be impacted. Always test for negative scenarios along with the positive ones. 5. Deploying a change in production should always be an incremental rollout. This means that initially the change will be effective in 1% of the production traffic. After certain bake time it increases to 10% and then if everything is fine eventually it will be effective across the entire production. 6. If any change requires manual intervention there should always be another person verifying your change before deploying. Another set of eyes drastically reduces manual errors. Mistakes are okay, but learning from them and building a resilient system to avoid any downtime is what matters at the end of the day. Feel free to comment your thought!! #SoftwareEngineer #Tech #Automation #Resilience #Learnings

170

12 Comments

How Pinterest built an end to end JSON logging system for clients apps ?? In early 2020, Pinterest realized that it didn’t have much visibility of how the app is running or a good system to look up for monitoring and troubleshooting. Problems with Logging - 1. All logs didn't have logging categories. 2. No good way to visualize/aggregate logs and set real time alerts. Key Design Decision while implementing new logging system -  1. Log payload should be schema-less and flexible, basically key-value pairs.  2. Logging service endpoint to handle logs validating, parsing, and processing. 3. Logs will be persisted in hive for any SQL-based queries. 4. A single and shared Kafka topic will be used for all logs going through this pipeline. 5. OpenSearch Integration for real time visualization and query tool.   Client side service integration will provide the metadata, and developers will just provide the name of the log and actual log payload. Feel free to add any relevant information in comments section and follow me for more such tech content. Reference - https://lnkd.in/gQA2QbBN #event #logging #opensearch #json #visualization #alerts #logs 

12