August 2026 Summaries

Coralogix describes how iGaming operators can connect user experience signals, operational telemetry, and business metrics to investigate revenue-affecting issues. In a demo scenario, product analytics identifies rage clicks on a deposit button, session replay reveals a player repeatedly attempting payment methods, and a correlated browser 402 error leads investigators through OpenTelemetry traces across payment, wallet, compliance, fraud, and acquirer services. Error logs ultimately identify a bank decline involving gambling merchants, while custom dashboards can measure the number of affected users and financial impact. The platform combines real user monitoring, session replay, logs, traces, metrics, and profiles, with source-side masking of sensitive data intended to support PCI DSS requirements. Coralogix also presents its AI assistant, Olly, as a faster way to analyze traces, estimate customer impact, and calculate costs, positioning observability as decision intelligence that links customer behavior, technical failures, and commercial outcomes.

Chris Cooney describes an automated guardrail-tuning loop for a fictional iGaming chat agent that can accept bets and provide market odds but must not predict market outcomes. The application uses an AWS Bedrock model, OpenTelemetry instrumentation, Coralogix Guardrails, and real-user monitoring to capture complete frontend-to-backend conversations in a telemetry lake. Claude Code runs Playwright-based tests through the actual user interface, evaluates whether policy-violating and permitted prompts are handled correctly, queries aggregated telemetry through the CX CLI, and iteratively updates guardrail rules and examples. Server-side aggregation and the CLI’s agent-oriented TOON format reduce token consumption, allowing many low-cost iterations while preserving model context. Over repeated tests, the system retained effective examples, discarded ineffective ones, and generalized rules beyond individual prompts, improving the chat interface’s ability to guard both inputs and outputs.

Zero-code OpenTelemetry instrumentation in Kubernetes can be implemented without the Instrumentation CRD by enabling Coralogix’s integrated auto-instrumentation through its Helm chart and using native pod-template annotations. Rather than maintaining separate custom resources, the OpenTelemetry Operator’s mutating webhook injects language-specific agents for Java, Python, .NET, or Node.js, while the inject-sdk annotation supplies OTLP environment variables for applications that already include OpenTelemetry libraries. This approach can reduce configuration drift, RBAC requirements, and upgrade-related schema maintenance while preserving standard Kubernetes manifests and GitOps workflows. Effective troubleshooting involves verifying annotations are placed on pod templates, confirming collector connectivity and API credentials, checking runtime support and language flags, avoiding double instrumentation, investigating failed init containers, and ensuring that no separate OpenTelemetry Operator webhook conflicts with the integrated installation.

Coralogix has integrated scheduled tasks for its AI agent Olly with the Notification Center, enabling recurring AI-generated reports to be delivered directly to Slack or email through existing label-based routing. Users can schedule hourly, daily, weekly, monthly, or one-time prompts for reports such as incident reviews that summarize root causes and unresolved cases, deployment digests based on operational event data, infrastructure health trend analyses comparing alerts and errors across periods, and AWS cost or FinOps reports. The feature builds on Olly’s existing case-analysis capabilities and Coralogix’s queryable telemetry, event, and billing data, aiming to automate operational reporting that teams often compile manually. Tasks are configured through the Olly AI settings by defining a prompt, model, recurrence, future start time, activation status, and optional delivery connectors, with guidance to specify the intended audience and report format.

Coralogix has introduced Product Analytics, an early-access capability that builds heatmaps, funnels, and pathways directly from the full-fidelity Real User Monitoring sessions already captured by its SDK, eliminating the need for a separate analytics SDK, tag manager, or duplicate behavioral-data pipeline. Heatmaps for web applications use Session Replay screenshots to show click, scroll, rage-click, and dead-click patterns, while funnels measure ordered user flows across web and mobile and can be segmented by attributes such as browser, country, account, environment, user, and app version. Pathways visualize the actual routes users take before or after a view, helping teams identify unexpected loops and exits. Each aggregate metric links back to individual sessions, Session Replay, frontend errors, and associated backend traces, logs, and metrics, allowing users to investigate conversion drops alongside their technical causes. Coralogix positions the feature as an extension of its Telemetry Data Lake, which unifies logs, metrics, traces, sessions, errors, and security events in customer-owned cloud object storage, and says it is available to eligible customers through account representatives or support.

Coralogix has introduced monthly token budgets for Olly API keys to help organizations manage the potentially unpredictable costs of autonomous AI investigations, which can consume substantially more tokens than standard chat interactions, particularly when multiple agents, tool calls, and planning steps are involved. The company argues that token use reflects investigative depth, since agents that inspect more logs, traces, and hypotheses may produce stronger results but incur higher costs. Olly is designed to reduce token requirements through its cloud-storage-based architecture, distributed DataPrime Query Engine, server-side aggregation, and metadata-driven query narrowing. In the Usage Management interface, administrators can establish a default monthly limit for all API keys or assign custom limits based on each key’s workload, while status indicators show whether usage is below 70% of a limit, approaching it, or has reached it. Coralogix recommends setting an initial default budget, observing usage over a billing cycle, and adjusting limits for keys that consistently support valuable automated work.

Coralogix announced a rebuilt Metrics Engine designed to support high-cardinality, high-dimensional metric workloads through a Rust-based distributed query system while retaining customer data in their own cloud object storage using the OpenTSDB format. Its new pricing model measures active time series received within the past 60 minutes rather than cumulative label combinations, aiming to avoid charges for redundant labels, higher sampling resolution, or additional dimensions. The release adds operational and historical storage tiers, with adaptive routing rules that move infrequently used metrics to lower-cost long-term storage and can return them to operational storage when usage rises; Coralogix says early migrations reduced metric footprint and spending by about 40 percent. It also introduces server-side deduplication for redundant high-availability collectors, triples default fair-use limits from 24 million to 72 million unique series per metric per day, and maintains compatibility with Prometheus remote-write, OTLP, PromQL, existing dashboards, alerts, and recording rules. The company presents the changes as particularly useful for AI agents and engineers who need richer labels, finer-grained data, and longer metric histories for production investigations.