<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Rahul Kumar on Medium]]></title>
        <description><![CDATA[Stories by Rahul Kumar on Medium]]></description>
        <link>https://medium.com/@rahulkumar49910?source=rss-3eedd4ac7c55------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*DcnsUcIsmdVC2ZalkDHh9A.jpeg</url>
            <title>Stories by Rahul Kumar on Medium</title>
            <link>https://medium.com/@rahulkumar49910?source=rss-3eedd4ac7c55------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Tue, 17 Nov 2026 03:42:09 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@rahulkumar49910/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Starting well in a new team: a manager’s guide ]]></title>
            <link>https://tech.loveholidays.com/starting-well-in-a-new-team-a-managers-guide-257b70be996d?source=rss-3eedd4ac7c55------2</link>
            <guid isPermaLink="false">https://medium.com/p/257b70be996d</guid>
            <category><![CDATA[leadership]]></category>
            <category><![CDATA[engineering]]></category>
            <category><![CDATA[culture]]></category>
            <dc:creator><![CDATA[Rahul Kumar]]></dc:creator>
            <pubDate>Mon, 21 Sep 2026 15:03:45 GMT</pubDate>
            <atom:updated>2026-09-21T15:04:58.043Z</atom:updated>
            <content:encoded><![CDATA[<p>At loveholidays, I recently moved teams as part of a wider organisational change. I spent time reflecting on past team changes I’ve been through and what it takes to succeed in this latest move. I’ve also drawn on the experiences of peers across tech to bring together a set of lessons for starting well in a new team. In this post, I use “manager” broadly to include lead engineers, engineering managers and team leads.</p><p>Moving teams as a manager can make you feel nervous or excited about the change. Both are valid reactions. Where you are in your career, where the team is in its lifecycle and how much change you’ve experienced before can all influence how you feel. Personally, I’m in the excited camp these days. Growth and innovation come from tackling unfamiliar problems and dealing with uncertainty. Nothing lasts forever, so every organisational shape and operating model is temporary.</p><p>The more I thought about the move, the more these six themes stood out: put people first, make manager handoffs explicit, establish a shared vision, use data to find focus, learn the new business domain and systems, and introduce change deliberately.</p><h4><strong>Put people first</strong></h4><p>When you join a new team, it’s easy to believe you need to demonstrate your knowledge immediately to prove yourself. That instinct is understandable, but it’s usually a trap. Remember that a manager’s role is to guide the team to success and make its output greater than the sum of its parts.</p><p>People are the most influential part of a team’s success, so building relationships and rapport with your new team should be your top priority. Spend time listening to how they see the team, where they see opportunities and what they want from their careers. Some people will share more than others, and some may not yet know what they want. That’s fine. Trust takes time, and people need space to work out their ambitions.</p><p>As you get to know each person, learn their strengths, development areas and ambitions. This will help you create opportunities that benefit both the individual and the team. For example, if an engineer is working towards a promotion, look for a high-visibility project they can lead, with the right support and clear expectations, aligned with their development goals.</p><h4><strong>Be intentional with manager handoffs</strong></h4><p>Being intentional about the manager handoff is another way to show care for your team, especially when you will be conducting the upcoming performance review cycle.</p><p><a href="https://larahogan.me/blog/manager-handoffs/">Lara Hogan has a great post about manager handoffs</a>. I’ve used this approach a few times now and believe it leaves everyone in a better place. In essence, this is a three-way meeting between the individual, their previous manager and their new manager to review objectives and key projects for the year, ensuring everyone has a shared understanding of where the individual is and what support they need.</p><p>Initially, the exercise may seem too formal, but you’d be surprised by how often people have things they’d like to clarify while they’re in the room with their old and new managers. This transparency is key to building healthy relationships and a culture of feedback.</p><h4><strong>Establish a shared vision</strong></h4><p>Building relationships helps you understand the team. Establishing a shared vision helps everyone understand where the team is heading and what matters most.</p><figure><img alt="A team all rowing in the same direction" src="https://cdn-images-1.medium.com/max/1024/1*ysbu3zg7JYiUbUFsaOqx2w.jpeg" /></figure><p>You can capture that vision in a lightweight format, such as a one-page charter or a few slides. It can clarify what matters most, help the team make trade-offs when priorities compete and show which parts of the team’s remit deserve the most attention. This gives people a basis for making better day-to-day decisions without waiting for direction.</p><p>Keep this at a high level: the goal isn’t to describe every detail of how the team works today, but to clarify where you want the team to be over the next few years.</p><p>Below is an example of what we did when forming our Checkout team back in 2024.</p><figure><img alt="Vision — To create a seamless and intuitive checkout experience, ensuring every purchase feels effortless and personalised" src="https://cdn-images-1.medium.com/max/916/1*9is9iE5DjUwyrhzuvab6dg.png" /></figure><figure><img alt="Mission — to expose a single checkout platform that provides customers with ease and choice, whilse capturing an exhaustive set of insights to drive informed decisions" src="https://cdn-images-1.medium.com/max/918/1*vt-T_XKgOAOGP7yZos2jkw.png" /></figure><p>If you are joining an existing team, some of this might already exist. If it does, use that and see whether it resonates with where you want to take the team. Even if it doesn’t initially, sit with it for a while rather than changing it immediately, then revisit it with the team in two to three months’ time.</p><h4><strong>Use data to find focus</strong></h4><p>With the people and vision in place, it’s time to understand how the team is performing as a system. Start by looking at the KPIs the team already tracks and considering how performance is changing over time. These might include measures such as conversion rates, page latency, errors, incidents, engagement, costs and performance. The important thing is to make the team’s health visible and create a shared understanding of how the team is performing against each measure.</p><p>In the Checkout team, we regularly reviewed funnel conversion rates, page latency, errors and other metrics in our stand-ups. Because we were familiar with what normal performance looked like, we noticed a regression that wasn’t triggering any corresponding alerts. Our app booking conversion rate had fallen following a change in a third-party system, even though the issue wasn’t producing errors in our systems. It was costing us around a dozen bookings a day, making the commercial impact clear and showing us that it needed to take priority over other initiatives.</p><p>If you join a team where meaningful KPIs don’t yet exist, work with the team to establish a small set of shared measures. They don’t need to be perfect to start with; the goal is to make the team’s health visible, visualise progress and support prioritisation conversations.</p><h4><strong>Learn the team’s business domain and systems</strong></h4><p>As a manager, you often need to wear both technical and product hats. You don’t need to know every implementation detail, but you should understand enough about the team’s work to recognise its challenges and know when to challenge an approach. You should also understand the problems the team is solving and why they matter to the business. Building relationships and trust with stakeholders will help you develop that context. You can model this for the team by showing them how to build and maintain those relationships themselves.</p><p>Pairing, collaboration, asking questions and meeting external stakeholders are great ways to learn from people. Observability tools and traces can help you understand how systems behave. And AI can accelerate the first pass by helping explain unfamiliar systems and trace how they fit together, but I’d still validate that understanding through independent investigation and conversations with the team.</p><p>This is also a good time to check what level of onboarding documentation the team has and whether you should invest in improving things for the next new joiner.</p><h4><strong>Evolution or revolution, take your time</strong></h4><p>Whether change is evolutionary or revolutionary, take time to understand the team before deciding how quickly to act. Moving too quickly to apply what worked in your previous team can create resistance and erode trust in you as the manager, even when the idea itself is sound. No one likes a manager who blindly introduces change simply because it worked for their last team.</p><p>Introduce ideas and ways of working gradually, and consult the team as you go. Share an idea in a few 1:1s to gauge an initial reaction. Use data or anecdotal evidence to explain why something might work. Run it as an experiment and check in with the team.</p><p>Sometimes, though, urgent problems require action. For example, you might join a team where the alerts channel is already an endless stream of noise. The team is experiencing alert fatigue: people have become so used to the noise that important alerts are being ignored, and incidents continue to occur. In such cases, use your experience and judgement to act deliberately, and clearly communicate why the change is needed. Even then, define how success will be measured and check the impact afterwards.</p><h4><strong>Wrapping up</strong></h4><p>Starting well in a new team is less about arriving with all the answers and more about creating the conditions for the team to succeed. Take time to listen, understand what matters and use evidence to decide where to focus. Then introduce change deliberately. You won’t get everything right immediately, but a thoughtful start gives the team a stronger foundation for what comes next.</p><p>If you’re moving teams now, or find yourself doing so in the future, use these ideas as prompts for how you want to approach the transition.</p><p><em>Shout out to Matt Koevort, Luke Williams, and Muammar Omar for helping shape the ideas here.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=257b70be996d" width="1" height="1" alt=""><hr><p><a href="https://tech.loveholidays.com/starting-well-in-a-new-team-a-managers-guide-257b70be996d">Starting well in a new team: a manager’s guide 📘</a> was originally published in <a href="https://tech.loveholidays.com/">loveholidays tech</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Stop Chasing Ghost Errors: A Systematic Approach to Frontend Error Monitoring That Actually Works]]></title>
            <link>https://tech.loveholidays.com/stop-chasing-ghost-errors-a-systematic-approach-to-frontend-error-monitoring-that-actually-works-fe5e1c8790cc?source=rss-3eedd4ac7c55------2</link>
            <guid isPermaLink="false">https://medium.com/p/fe5e1c8790cc</guid>
            <category><![CDATA[frontend]]></category>
            <category><![CDATA[software-engineering]]></category>
            <dc:creator><![CDATA[Rahul Kumar]]></dc:creator>
            <pubDate>Tue, 22 Jul 2025 08:16:58 GMT</pubDate>
            <atom:updated>2025-07-22T17:12:36.208Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="Man cornering ghost with vacuum" src="https://cdn-images-1.medium.com/max/832/1*RDEN-mIu5IlBKkUHgnLKZQ.jpeg" /></figure><p>Every frontend team knows the frustration: <a href="https://sentry.io/welcome/">Sentry</a> alerts flooding in, stack traces pointing to code you’ve never seen, and the nagging question of whether these errors actually impact users. The challenge isn’t just monitoring errors — it’s separating the signal from the noise in an environment where your code runs alongside browser extensions, third-party scripts, and countless variables beyond your control.</p><p>As frontend engineers we need to be aware that our application’s runtime environment can be affected by:</p><ul><li>Varying device capabilities, screen sizes, performance and network constraints</li><li>Different browsers and their rendering engines</li><li>Browser extensions like Honey, MetaMask, or ad blockers</li><li>Third-party integrations and scripts loaded outside your application code (like Google Tag Manager), all of which may not be owned by the same team</li></ul><p>This complexity creates a perfect storm for cryptic, hard-to-debug errors that leave engineers questioning whether they’re chasing real problems or environmental ghosts.</p><p>At loveholidays, we’ve developed a systematic approach to ensure our error monitoring tools surface actionable insights, rather than confusing noise. Since implementing this framework, we’ve reduced our 30 day rolling average of errors for our primary frontend application by 70%.</p><h3>Measure your baseline</h3><p>Lord Kelvin once said “If you can’t measure it, you can’t improve it” — advice that’s particularly relevant when tackling error rates.</p><p>Start by setting up visualisations that track your total error counts over time. This gives your team and stakeholders clear visibility into trends and creates a satisfying feedback loop as engineers watch those numbers drop.</p><p>Beyond motivation, this baseline becomes crucial for measuring the impact of your improvements and catching regressions before they spiral out of control.</p><p>At loveholidays, we’ve established a top-level error KPI dashboard that enables engineers to analyse errors over time by team and service. As shown below, we use a 30-day rolling average to visualise long-term progress, including year-over-year comparisons.</p><figure><img alt="30 day average of errors" src="https://cdn-images-1.medium.com/max/1024/0*bPXTC1_bINIYxvsP" /></figure><p>Beyond service-level metrics, we needed to understand errors at per session level, so we created a dedicated view for that perspective.</p><figure><img alt="Errors per session graph trending downwards" src="https://cdn-images-1.medium.com/max/1024/0*FHiktKAFCtYa2sab" /></figure><p>Together, these dashboards provide the visibility needed to track our error reduction efforts and validate that our engineering investments are delivering meaningful improvements to user experience.</p><h3>Capture more context</h3><p>Default error monitoring typically captures the basics — browser, device type, release version, and stack trace. But adding more specific contextual information can change cryptic errors to become far more actionable.</p><h4>Essential application specific context to add</h4><p>A few essential bits of info at loveholidays to complement the defaults from Sentry are</p><ul><li><strong>Render mode</strong> — this allows us to identify whether errors happened during server-side or client-side rendering, as this significantly affects debugging approaches. Alternatively you can create separate projects for these in your error monitoring tool to track these in complete isolation.</li><li><strong>Modern or legacy build </strong>— If you serve different builds based on browser capabilities, knowing which build generated an error is crucial. When modern JavaScript features accidentally slip into legacy builds, this context makes the problem immediately obvious.</li><li><strong>Breadcrumbs</strong> — Capture the sequence of actions leading to errors. Understanding that a user “navigated from checkout to payment, then the error occurred” provides invaluable debugging context.</li></ul><h4>Session replay links</h4><p>Tools like FullStory or Hotjar capture user session recordings that Product and Analytics teams often use for insights. If you’re already using these tools, integrate them with your error monitoring so engineers can jump from error reports directly to watching what the user experienced. This can help engineers quickly triage if an error has customer impact and if so, what exactly happens.</p><p>At loveholidays we use <a href="https://www.fullstory.com/">Fullstory</a> and <a href="https://sentry.io/">Sentry</a>, so here’s how we connect them:</p><figure><img alt="Code snippet of adding session replay links to errors" src="https://cdn-images-1.medium.com/max/1024/1*OKW1D0dVwTbvZo1TfIQfnw.png" /></figure><h4>Detect browser translation</h4><p>Browser translation is incredibly helpful for users, but it can break frontend code in unexpected ways (<a href="https://github.com/preactjs/preact/issues/948">Preact rendering, for example</a>). If your application has a default or selected locale, detecting if translation is applied is straightforward — just compare the document’s lang attribute to your expected locale:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*YG6nRRM4EIDV82jzjCSU2Q.png" /></figure><h4>Identify the true source of an error</h4><p>Not all errors are created equal. Stack traces from your application code deserve different treatment than errors from browser extensions or third-party scripts. The key is automatically categorising errors so engineers can focus their time appropriately.</p><p>Here’s a typical example of a third-party error that may appear in your monitoring:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/863/0*H9rYhNbfYTYnqdF6" /></figure><p>Errors like this need different handling than application bugs — you might ignore them, contact the integration partner, or implement workarounds.</p><p>At loveholidays, we use a two-pronged approach:</p><p><strong>First</strong>, we include Sentry’s Vite and Webpack plugins in our build pipeline, using their <a href="https://docs.sentry.io/platforms/javascript/configuration/filtering/#using-thirdpartyerrorfilterintegration">third party error filtering to drop errors</a> where all frames come from external code. When we introduced this, we immediately filtered out about 20% of our main frontend application’s errors.</p><p><strong>Second</strong>, we have custom logic that identifies specific sources and tags errors accordingly:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*1nyXonY5fVA-1PmDSeInaQ.png" /></figure><h3>Take action</h3><p>With rich context and proper categorisation, your team can make informed decisions about error prioritisation. Engineering is about tradeoffs, and sometimes the right choice is to ignore errors that have zero customer impact and are beyond your control — such as errors occurring within third-party scripts that offer no clear path to resolution. Where possible, consider raising issues with the relevant vendor or open-source project so these problems can eventually be addressed upstream for the wider community.</p><p>Our <a href="https://docs.sentry.io/platforms/javascript/configuration/filtering/#using-ignore-errors">Sentry ignore list</a> has grown over the years, and that’s perfectly fine. The goal isn’t zero errors — it’s <a href="https://tech.loveholidays.com/from-firefighting-to-foresight-a-new-culture-of-software-reliability-affcf565f9b3">zero noise and surprises</a> preventing you from seeing real problems.</p><p>For genuine issues, we ensure teams can actually address them by building error-fixing time into our regular processes. Platform improvements get prioritised alongside “feature” work, and on-call engineers have dedicated time to investigate and resolve monitoring alerts without the pressure of feature delivery.</p><p>Improvement without sustainability is just temporary relief. As your error counts drop, set up systems or processes to catch regressions early — either through automated alerting on error volume spikes or regular team reviews of error trends.</p><h3>Conclusion</h3><p>By implementing these techniques systematically, we’ve elevated our error monitoring from helpful but noisy, to a finely-tuned system that surfaces exactly what needs attention. At loveholidays, we’ve reduced our error volumes by 70% whilst dramatically increasing engineer confidence in our monitoring.</p><p>The key insight is that in the frontend world, not all errors are created equal. Some deserve immediate attention because they impact users directly, whilst others can be safely ignored because they’re outside your control and we’re confident they do not impact the user. And many fall somewhere in between where context determines the right action.</p><p>When your error monitoring shows genuine application issues rather than environmental noise, every alert becomes an opportunity to improve user experience. Your engineering team can focus on building features and fixing real problems instead of deciphering cryptic error messages from browser extensions or third-party scripts.</p><p>The investment in better error monitoring pays dividends: faster debugging, clearer production understanding, and ultimately happier users experiencing fewer issues. Start with measuring your baseline, enrich your errors with meaningful context, and watch as your team transforms from hunting down vague issues to confidently tackling the specific problems that impact users.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=fe5e1c8790cc" width="1" height="1" alt=""><hr><p><a href="https://tech.loveholidays.com/stop-chasing-ghost-errors-a-systematic-approach-to-frontend-error-monitoring-that-actually-works-fe5e1c8790cc">Stop Chasing Ghost Errors: A Systematic Approach to Frontend Error Monitoring That Actually Works</a> was originally published in <a href="https://tech.loveholidays.com/">loveholidays tech</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>