QA Engineer - O2C & Enterprise Billing Testing

givestar selects New Relic to boost platform reliability and help 250,000+ charities maximize donations. Fast-growing fundraising platform aims to accelerate root cause analysis, minimise mean-time-to-resolution, and build AI-driven, self-healing systems as it expands across the UK and US September 3, 2026 LONDON, UK - 3 September, 2026 - New Relic, the Intelligent Observability company, announced that givestar, a rapidly-growing UK-based charity-technology platform, has chosen New Relic as its observability partner. givestar plans to leverage New Relic's Intelligent Observability Platform, including its Application Performance Monitoring (APM), Infrastructure Monitoring, and AI capabilities. This will enable givestar to boost the resilience of its systems, eliminate manual log searching, and ensure seamless experiences for over 750,000 signed up users, millions of donors and more than 250,000 non-profit organizations. Prior to deciding to adopt New Relic, givestar relied on out-of-the-box Google Cloud Platform (GCP) monitoring tools. However, with its own platform scaling rapidly - adding 10,000 new users every week - the engineering team recognised the need for an observability platform to speed up root cause analysis (RCA) and maintain seamless performance during traffic peaks. By migrating to New Relic, givestar gains end-to-end telemetry across its application stack and infrastructure. Customised dashboards give developers, QA, product managers, and technical leadership real-time visibility into application health, ensuring the team identifies potential issues before they ever reach end users. "For charities, maximising the amount of money raised is absolutely critical. Having a resilient system that gives a brilliant experience to donors directly impacts how much support charities get," said givestar CTO Mehdi Shahin. "Our ambition is to be AI-first. New Relic stands out because it's exceptionally developer-friendly, cuts our incident investigation time through superior RCA, and provides the AI roadmap we need to move towards self-healing systems." Key implementation highlights and objectives: * Accelerating RCA and lowering MTTR: By replacing fragmented log searches with New Relic's developer-friendly APM and AI tools, givestar aims to dramatically reduce its mean time to resolution (MTTR) and identify post-deployment issues instantly. * Proactive two-layer alerting: givestar will introduce a multi-tier alerting strategy to distinguish critical night-time escalations from lower-priority indicators. This ensures engineers intervene long before performance degradations impact live fundraising campaigns. * Scaling for peak events: As givestar scales its platform integrations, partnerships, and US expansion, New Relic's combined infrastructure and application dashboards give the entire team a shared, real-time command center during high-volume traffic peaks. * AI and self-healing ambitions: givestar plans to integrate New Relic's agentic AI features directly into its internal orchestration systems to automate incident diagnosis and remediation. "Behind every donation processed on givestar is a cause that relies on platform reliability. Fast-growing scale-ups like givestar simply cannot afford unexpected downtime or performance slowdowns during crucial fundraising events," said New Relic SVP and GM EMEA David Cruddas. "Equipping their team with intelligent observability means potential issues are identified and resolved before they're noticed by donors or charities. We're incredibly proud to partner with givestar on their AI journey and as they build the technical foundation needed to power their next phase of international expansion." New Relic will serve as a foundational building block for givestar as it scales its operations across the UK and the US markets. About givestar givestar is the next generation fundraising platform helping people raise more for charity through smarter technology, better storytelling, and community powered campaigns. Built to make giving feel simple and native in the places people already spend time, givestar supports fundraisers taking on everything from first time challenges to limit-pushing endurance feats. givestar, which is a proud B Corp, has processed more than £50m in donations since its inception and is attracting more than 10,000 new sign ups each week. About New Relic. New Relic arms businesses with the trust and confidence required to thrive in the AI era. The New Relic Intelligent Observability Platform is the leading AI-strengthened platform designed to unify telemetry and business outcomes, bringing intelligence and automated actions to the most complex digital environments. The platform shifts teams from reactive firefighting to intelligent orchestration, leveraging AI-driven automation to optimize technology spend and protect revenue in real-time. That's why global leaders - Adidas Runtastic, Domino's, Ryanair, Swiggy, Topgolf, and William Hill - run on New Relic to drive innovation and deliver exceptional customer experiences. Visit: www.newrelic.com. Media contact. New Relic, Inc.

AI observability must evolve for the agentic era. 2026-09-02 15:09 Read the original article: Grafana Labs has announced the general availability of six AI capabilities, extending Grafana Assistant into an agentic operations layer that detects, investigates, and remediates production issues. The releases include Grafana Assistant Investigations, Grafana Assistant Workspace, Grafana Assistant Automations, the Grafana Cloud MCP server, gcx, and Grafana Agent Observability. Observability has... Security Products & Services July 28, 2026 New Relic has announced AI Coding Observability, an open-source tool for monitoring AI-assisted software development workflows. As organizations adopt AI coding assistants, these tools often operate outside existing observability systems, limiting visibility into their use. AI Coding Observability extends monitoring into the software development process, enabling organizations to track, analyze,... June 8, 2026 Security researchers have disclosed "GhostJacking," a new class of attacks that exploits trusted observability and security platforms to manipulate AI... August 11, 2026

Magento monitoring vs APM: 'is it fast?' is not 'is it working?' 2026-08-20 · Byte8 Team Every "best Magento monitoring tools" listicle names the same four: New Relic, Datadog, Blackfire, Tideways. They're all excellent - and they all answer the same question: *is the site fast?* That's application performance monitoring, and it's essential. But "is it fast?" is not the same job as "is it working?" - whether a customer can browse, search, and buy right now, and whether the money is actually landing. A Magento store can be lightning-fast and quietly broken at the same time, and your APM dashboard will stay green through every second of it. Magento monitoring vs APM: what's the difference? Application performance monitoring (APM) measures how the system performs: response time, throughput, database time, memory, and the stack traces behind slow or failing requests. It answers "is it fast, and is the code healthy?" Business-outcome monitoring measures whether the store's actual job is getting done: can a customer complete checkout, is search returning products, is content un-tampered, are orders and payments flowing. It answers "is it working, and is it still making money?" Both are monitoring. They watch different layers, catch different failures, and neither substitutes for the other. The trap is owning only the first and believing a green performance dashboard means a healthy store. What APM and profilers are genuinely great at. Be fair to the four names above, because they're very good at their job. * New Relic and Datadog are APM: distributed tracing, database query time, throughput, PHP exceptions, infrastructure metrics. When checkout is slow, they show you the N+1 query, the saturated connection pool, or the third-party call adding 600ms. * Blackfire and Tideways are profilers: they walk a single request and point at the exact function eating 800ms, so a performance fix lands on the real cause instead of a guess. If your TTFB is creeping, a page got heavy, or the database is the bottleneck, these are the correct tools and nothing here replaces them. But notice what they're built to answer: a *performance* question. A failure that isn't slow and doesn't throw an exception sits, by design, outside their field of view. Why a fast Magento store can still be broken. Here are four failures that are fast, green, and exception-free - and still cost you customers: * A 200-but-broken checkout. A JavaScript regression after a deploy breaks the "Place Order" button. The server happily serves 200 OK for every checkout page; no customer can actually pay. APM sees fast, healthy responses. * An injected card skimmer. A few lines of JavaScript are slipped into a CMS block or a "Miscellaneous HTML" config field. They load on every page in 400ms and copy card details at checkout. Fast, cached, exception-free - and quietly stealing cards. * A drifted indexer. The category indexer falls into an invalid state and product listing pages render zero products. Every page returns 200 in good time. Customers land in an empty shop. * An uncaptured payment. An order is authorised at checkout, goes to fulfilment, and ships - but the *capture* that actually collects the money silently fails. Nothing errors, nothing is slow. You find it weeks later, reconciling the bank against your orders. None of these is a performance problem. None throws an exception. Each one returns HTTP 200. That is *precisely* why an APM dashboard stays green - and why "is it fast?" can't protect revenue on its own. The failures split neatly by whether a performance tool can even see them: failure slow? throws? HTTP APM sees it? - - - - - broken "Place Order" (JS) no no 200 no injected card skimmer no no 200 no drifted indexer / empty PLP no no 200 no uncaptured payment no no 200 no N+1 query on checkout YES no 200 yes unhandled PHP exception maybe YES 500 yes APM catches the bottom two beautifully. The top four are invisible to it - not through misconfiguration, but because they're the wrong shape for a performance tool to detect. The layer APM can't reach. To answer "is it working?", you need two things a trace can't give you: a real browser, and a way to read Magento's own state. From the outside - a real browser. Synthetic checks load the store in real Chromium, not a HEAD request, and walk the funnel: open a product, add to cart, go to cart, enter shipping, reach payment. Each step is timed independently and the browser captures console.error and unhandled exceptions along the way. When a deploy breaks add-to-cart, the check fails at that exact step and names the JavaScript error - instead of a vague "site down" the next morning. From the inside - Magento's own state. Some failures are upstream of the page: an indexer mid-drift, a cron about to cause a backlog, OpenSearch gone red, a skimmer injected into a CMS block. A read-only health endpoint surfaces them - Pulsar's exposes 20 collectors, each mapped to one silent failure: indexer state, cron heartbeat, queue depth, OpenSearch health, stuck pending_payment orders, content integrity, SSL expiry, admin 2FA coverage, and more. This is the layer Pulsar occupies. It doesn't profile your code or trace your queries - that's what your APM is for. It watches the business outcome. A note on honesty, because this category is still maturing: a broken checkout, a drifted indexer, and an injected skimmer are all things you can watch continuously *today*. Others - reconciling that every authorised order was actually captured, for instance - are harder problems that monitoring is still growing into. "Is it working?" is a bigger question than any single tool has fully answered, and pretending otherwise helps no one. APM vs business-outcome monitoring, side by side. * What it watches: APM - the system (traces, DB time, memory, exceptions). Business-outcome - the funnel and Magento's state (checkout, search, content, orders). * The question it answers: APM - *"is it fast?"* Business-outcome - *"is it working?"* * The failure it's built to catch: APM - a slow query, a saturated server, a thrown exception. Business-outcome - a 200-but-broken checkout, a live skimmer, a drifted indexer. * Its blind spot: APM - anything that's fast, green, and exception-free. Business-outcome - a slow-but-working store (that's APM's job, not this layer's). * When it pages you: APM - latency and error-rate thresholds. Business-outcome - a real customer action stops working. Common questions. Is APM enough to monitor a Magento store? No. APM tells you whether the store is fast and whether the code is throwing errors - necessary, but not sufficient. The failures that most directly cost revenue (a broken "Place Order" button, an injected skimmer, a drifted indexer) are fast and exception-free, so they never reach an APM dashboard. You need a layer that checks the business outcome as well. Does business-outcome monitoring replace New Relic or Datadog? No - it complements them. New Relic and Datadog answer performance and infrastructure questions nothing else answers as well. Pulsar answers "can a customer still buy?", a different question on a different layer. Run both; they catch different failures. What can synthetic checkout monitoring catch that APM can't? A synthetic check drives the real funnel in a real browser, so it fails the moment a customer *can't* complete a step - even when every underlying HTTP response is a fast 200. A JavaScript error that breaks add-to-cart, a payment step that no longer renders, a shipping method that vanished after a deploy: invisible to a trace, caught by a browser that actually tries to buy. Where do profilers like Blackfire and Tideways fit? They're for performance work, not availability. Once you've found a slow page - often *via* APM or a synthetic timing - a profiler tells you which function to fix. They answer "why is this slow?", not "is the store working?". You don't choose - you layer. This was never "replace New Relic with Pulsar." Keep your APM; it does a job nothing else does, and a fast store is genuinely worth engineering for. The mistake is stopping there and reading a green performance dashboard as proof the store is healthy. Add the layer that watches the outcome - the funnel, the content, the internal state - and "the store is up" finally starts to mean "customers can buy." Fast is table stakes. Working is the revenue. Enjoyed this? Share it with your team.