{"id":744,"date":"2026-07-05T05:00:29","date_gmt":"2026-07-05T05:00:29","guid":{"rendered":"https:\/\/struct.ai\/articles\/watchdog-rca-automated-root-cause\/"},"modified":"2026-07-05T05:00:29","modified_gmt":"2026-07-05T05:00:29","slug":"watchdog-rca-automated-root-cause","status":"publish","type":"post","link":"https:\/\/struct.ai\/articles\/watchdog-rca-automated-root-cause\/","title":{"rendered":"Watchdog RCA: Automated Root Cause Analysis Explained"},"content":{"rendered":"<p><em>Written by: Nimesh Chakravarthi, Co-founder &amp; CTO, Struct<\/em><\/p>\n<h2 id=\"key-takeaways\">Key Takeaways for On\u2011Call Teams<\/h2>\n<ul>\n<li>Watchdog RCA tools correlate metrics, logs, traces, and code changes to surface probable root causes within minutes of an alert.<\/li>\n<li>Manual triage typically wastes 45\u201360 minutes of MTTR, and coordination overhead plus tribal knowledge gaps slow growing teams even more.<\/li>\n<li>Datadog Watchdog RCA is confined to Datadog data, lacks Slack-native delivery, and cannot correlate external code changes or runbooks.<\/li>\n<li>Proactive automated RCA platforms run cross-stack investigations, build unified timelines, and summarize blast radius directly in existing communication channels.<\/li>\n<li>Struct automates your on-call runbook and delivers complete root cause hypotheses in under five minutes, so Seed-to-Series C teams avoid manual log-hunting.<\/li>\n<\/ul>\n<h2>The Problem: Manual Triage Slows Growing Engineering Teams<\/h2>\n<p>At 3 AM, an alert fires and a senior engineer opens five browser tabs: Datadog, AWS CloudWatch, Sentry, GitHub, and a stale runbook doc. They start correlating malformed log lines by hand while the incident clock keeps running. A large share of MTTR for production incidents disappears into this diagnosis and root cause analysis phase. For Seed-to-Series C teams bound by SLAs, every minute in that window increases risk.<\/p>\n<p>The structural problem compounds as teams and systems grow. Most SRE teams see median P1 MTTR between 45\u201360 minutes without automation, with roughly 12 of those minutes lost to coordination overhead alone. That overhead compounds when manual runbooks rely on tribal knowledge; when the expert is unavailable, new SREs lack the critical procedural context to proceed. The financial stakes are severe: <a href=\"https:\/\/coralogix.com\/guides\/alert-fatigue\" target=\"_blank\" rel=\"noindex nofollow\">over 90% of midsize and large enterprises report that a single hour of downtime costs more than $300,000<\/a>, so every minute of delay from coordination friction or missing context directly erodes margin while alert fatigue keeps SLA breach timers running.<\/p>\n<h2>The Solution Category: Automated Root Cause Analysis Tools<\/h2>\n<p>Automated root cause analysis tools replace the manual correlation phase of incident response with machine-driven investigation. Instead of waiting for an engineer to query each data source, these platforms ingest alert streams, observability telemetry, deployment events, and code context at the same time, then surface a ranked hypothesis before a human gets involved.<\/p>\n<p><a href=\"https:\/\/doi.org\/10.36227\/techrxiv.177126526.67932591\/v1\" target=\"_blank\" rel=\"noindex nofollow\">Root cause analysis (RCA) in cloud microservice systems remains a critical challenge, with Mean Time to Identify Root Cause often exceeding 60 minutes<\/a>, while traditional manual reconstruction can take hours or days. Teams shorten MTTR by automating the manual correlation work that currently dominates that window.<\/p>\n<h2>How Automated RCA Platforms Deliver Faster Incident Response<\/h2>\n<p>Teams achieve these MTTR gains through a specific set of capabilities that replace manual triage steps with automated ones.<\/p>\n<ul>\n<li><strong>Zero-click investigation:<\/strong> Alerts trigger automated analysis without engineer input.<\/li>\n<li><strong>Unified timeline:<\/strong> Logs, traces, metrics, and deploy events appear in one consolidated view.<\/li>\n<li><strong>Blast radius assessment:<\/strong> The platform immediately scopes affected users or services.<\/li>\n<li><strong>Runbook encoding:<\/strong> Custom operational procedures sit inside the investigation logic.<\/li>\n<li><strong>Communication-channel delivery:<\/strong> Results surface in Slack or PagerDuty threads where incidents already live.<\/li>\n<li><strong>Handoff to remediation:<\/strong> Root cause context flows directly to coding agents or PR creation.<\/li>\n<\/ul>\n<p><a href=\"https:\/\/hud.io\/blog\/automated-root-cause-analysis-lowers-mttr\" target=\"_blank\" rel=\"noindex nofollow\">Effective automated RCA integration correlates metrics, logs, traces, and change events to highlight the single most likely starting point, attaching ownership context such as on-call team, runbooks, and dashboards to turn alerts into response plans<\/a>.<\/p>\n<h2>Manual, Assisted, and Automated RCA: A Maturity Progression<\/h2>\n<p>Incident response maturity usually progresses through three stages, from fully manual work to proactive automation.<\/p>\n<p><strong>Manual:<\/strong> The engineer queries each tool independently, correlates by memory, and writes a postmortem from scratch. Manual triage forces engineers to reconstruct the incident timeline from the ground up every time, which recreates the 60\u201390 minute post-mortem burden mentioned earlier.<\/p>\n<p><strong>Assisted (AI chatbot):<\/strong> The engineer pastes logs into ChatGPT or Claude and prompts iteratively. The work remains reactive because the engineer still wakes up, gathers data, and guides the model. Context window limits and malformed log formatting further degrade output quality under pressure.<\/p>\n<p><strong>Automated (proactive platform):<\/strong> The investigation starts the moment the alert fires. LLM-powered root cause analysis produces a complete RCA draft significantly faster than a manual version that would have taken hours. The engineer reviews and refines a finished summary instead of building one from scratch. The strategic decision shifts from whether to automate to when incident volume and SLA pressure justify that step.<\/p>\n<h2>Implementation Stages and Operational Maturity<\/h2>\n<p><a href=\"https:\/\/hud.io\/blog\/automated-root-cause-analysis-lowers-mttr\" target=\"_blank\" rel=\"noindex nofollow\">A practical adoption plan for automated RCA begins with instrumenting one \u201cgolden\u201d incident path, such as checkout, login, or payment confirmation, and building a minimum signal set of alerts, traces, deploy or change events, topology, and ownership or runbook links<\/a>. Teams then expand coverage incrementally across services and failure modes. Maturity indicators include earlier cause detection, fewer war-room participants, faster handoffs, and lower total alert volume per incident.<\/p>\n<p>Some RCA tools require weeks of data collection and ML model training before providing useful analysis, while others can investigate incidents from day one by querying existing data sources. Time-to-value becomes the critical differentiator for fast-moving engineering teams that cannot pause delivery for a long training period.<\/p>\n<p><a href=\"https:\/\/cal.com\/deepanm\/struct-demo\" target=\"_blank\"><strong>See how Struct delivers immediate time-to-value<\/strong><\/a> by connecting your stack in under 10 minutes and running your first automated investigation the same day.<\/p>\n<h2>Datadog Watchdog RCA in Practice<\/h2>\n<p>Datadog Watchdog is an AI-powered anomaly detection layer built into the Datadog platform. Watchdog RCA extends this layer by attempting to identify a root cause for anomalies it detects, correlating metric changes, APM traces, and log patterns within the Datadog ecosystem. It surfaces findings inside the Datadog UI and can trigger monitors or send notifications to connected channels.<\/p>\n<p>Teams already heavily invested in Datadog gain a starting point for investigation without leaving the platform. Watchdog RCA works best for single-service anomalies where all relevant telemetry already flows through Datadog and cross-stack context is less critical.<\/p>\n<h2>Where Datadog Watchdog RCA Falls Short for Complex On\u2011Call<\/h2>\n<p>Several structural constraints limit Watchdog RCA\u2019s effectiveness in complex, multi-service production environments.<\/p>\n<ul>\n<li><strong>Ecosystem lock-in:<\/strong> Watchdog RCA only reasons over data inside Datadog. Teams using AWS CloudWatch, GCP Logs, Sentry, or Azure Traces for parts of their stack receive incomplete analysis. The wrapper problem occurs when AI lacks access to external tools such as GitHub, Jira, or PagerDuty and relies only on available data, preventing it from answering \u201cwhat changed?\u201d without explicit manual input.<\/li>\n<li><strong>Correlation vs. causation:<\/strong> Correlation-based AI RCA tools identify temporal correlations between events but fail to distinguish them from true causation, which is the primary failure mode in incident management AI.<\/li>\n<li><strong>Stale runbook context:<\/strong> AI RCA tools suffer from stale context by suggesting fixes based on outdated runbooks unless the team continuously maintains those runbooks inside the platform.<\/li>\n<li><strong>No Slack-native delivery:<\/strong> Watchdog findings surface in the Datadog UI, so engineers must context-switch away from their incident communication channel to review analysis.<\/li>\n<li><strong>No code context:<\/strong> Watchdog RCA does not natively correlate anomalies with GitHub commits, pull requests, or exception traces from Sentry, which leaves the \u201cwhat changed in the code?\u201d question unanswered.<\/li>\n<li><strong>Alert overload without deduplication:<\/strong> Major incidents can trigger a high volume of alerts, many of which are duplicates or symptoms rather than root causes. Watchdog surfaces anomalies but does not consolidate them into a single incident narrative with ranked hypotheses.<\/li>\n<\/ul>\n<h2>Watchdog RCA vs. Struct for 3 AM Incidents<\/h2>\n<p>The table below compares Datadog Watchdog RCA and Struct on four dimensions that matter directly for on-call triage. All figures come from cited sources or vendor documentation.<\/p>\n<table>\n<thead>\n<tr>\n<th>Dimension<\/th>\n<th>Datadog Watchdog RCA<\/th>\n<th>Struct<\/th>\n<th>Source<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td><strong>Investigation time<\/strong><\/td>\n<td>Surfaces anomaly correlation, with full triage still manual (industry baseline: 45\u201360 min P1 MTTR)<\/td>\n<td><a href=\"https:\/\/www.producthunt.com\/products\/struct-2\" target=\"_blank\">Under 5\u201310 minutes, automated before engineer opens laptop; 80% triage reduction (45 min \u2192 5 min)<\/a><\/td>\n<td>incident.io; <a href=\"https:\/\/www.producthunt.com\/products\/struct-2\" target=\"_blank\">Product Hunt<\/a><\/td>\n<\/tr>\n<tr>\n<td><strong>Setup time<\/strong><\/td>\n<td>Included in Datadog, but requires full Datadog instrumentation<\/td>\n<td><a href=\"https:\/\/www.producthunt.com\/products\/struct-2\" target=\"_blank\">10-minute setup; connects to existing observability stack<\/a><\/td>\n<td><a href=\"https:\/\/www.producthunt.com\/products\/struct-2\" target=\"_blank\">Product Hunt<\/a><\/td>\n<\/tr>\n<tr>\n<td><strong>Primary interface<\/strong><\/td>\n<td>Datadog UI, which requires a context-switch from Slack<\/td>\n<td>Slack-native, with a dynamic dashboard linked from the alert thread<\/td>\n<td>Vendor documentation<\/td>\n<\/tr>\n<tr>\n<td><strong>Investigation approach<\/strong><\/td>\n<td>Correlation within Datadog telemetry only<\/td>\n<td><a href=\"https:\/\/www.producthunt.com\/products\/struct-2\" target=\"_blank\">Cross-stack: metrics, logs, traces, code, and custom runbooks<\/a><\/td>\n<td><a href=\"https:\/\/www.producthunt.com\/products\/struct-2\" target=\"_blank\">Product Hunt<\/a><\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p><strong>A 3 AM investigation timeline with Struct:<\/strong><\/p>\n<ol>\n<li><strong>T+0:00<\/strong>, alert fires in the Slack alerting channel.<\/li>\n<li><strong>T+0:30<\/strong>, Struct automatically queries logs, metrics, traces, and GitHub.<\/li>\n<li><strong>T+3:00<\/strong>, Struct posts a blast radius summary, including affected users and services, directly in the Slack thread.<\/li>\n<li><strong>T+5:00<\/strong>, a full root cause hypothesis, supporting evidence, and suggested fix appear in a dynamic dashboard linked from Slack.<\/li>\n<li><strong>T+6:00<\/strong>, the engineer wakes up, reads the summary, and confirms or challenges the hypothesis through a conversational Slack bot.<\/li>\n<li><strong>T+8:00<\/strong>, the engineer hands off the confirmed root cause to a coding agent or generates a PR fix directly from Struct.<\/li>\n<\/ol>\n<h2>Which RCA Tools Deliver the Strongest Impact?<\/h2>\n<p>The table below shows how different RCA tool categories perform across three critical dimensions: impact on MTTR, ability to reduce alert noise, and breadth of data sources they can query. Proactive Slack-native AI platforms deliver the largest MTTR reduction because they remove the reactive bottleneck entirely, while ecosystem-native tools like Watchdog RCA improve efficiency only within their own platform boundaries.<\/p>\n<table>\n<thead>\n<tr>\n<th>Tool Category<\/th>\n<th>MTTR Impact<\/th>\n<th>Alert Noise Reduction<\/th>\n<th>Cross-Stack Coverage<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>Proactive Slack-native AI (e.g., Struct)<\/td>\n<td><a href=\"https:\/\/www.producthunt.com\/products\/struct-2\" target=\"_blank\">80% triage reduction; 45 min \u2192 5 min<\/a><\/td>\n<td>Automated deduplication and transient versus critical classification<\/td>\n<td>Logs, metrics, traces, code, and custom runbooks<\/td>\n<\/tr>\n<tr>\n<td>AIOps platforms (e.g., Dynatrace, Selector)<\/td>\n<td><a href=\"https:\/\/aiopsschool.com\/blog\/mastering-aiops-for-root-cause-analysis-best-practices-for-modern-it-operations\" target=\"_blank\" rel=\"noindex nofollow\">MTTR from hours to minutes<\/a><\/td>\n<td><a href=\"https:\/\/selector.ai\/learning-center\/aiops-tools-key-features-and-top-8-solutions\" target=\"_blank\" rel=\"noindex nofollow\">70\u201390% alert volume reduction<\/a><\/td>\n<td>Broad coverage with enterprise setup required<\/td>\n<\/tr>\n<tr>\n<td>Ecosystem-native (e.g., Datadog Watchdog RCA)<\/td>\n<td>Partial improvement that reduces manual query time within one platform<\/td>\n<td>Within-platform deduplication only<\/td>\n<td>Limited to ingested Datadog data<\/td>\n<\/tr>\n<tr>\n<td>Reactive AI chatbot (Claude or ChatGPT)<\/td>\n<td>Marginal impact because the engineer still gathers all data manually<\/td>\n<td>No automated reduction<\/td>\n<td>Restricted to what the engineer pastes in<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<h2>Prerequisites, Risks, and Edge Cases<\/h2>\n<p>Automated RCA tools perform best when the underlying observability stack is instrumented consistently. <a href=\"https:\/\/hud.io\/blog\/automated-root-cause-analysis-lowers-mttr\" target=\"_blank\" rel=\"noindex nofollow\">Correlation-based automated RCA tools suffer from garbage-in-garbage-out failures when signals lack consistent service tags, use inconsistent naming conventions, or omit trace propagation across async boundaries<\/a>.<\/p>\n<p>Teams with minimal logging, no trace IDs, or no alerting triggers will not receive accurate automated investigations from any tool. The baseline prerequisites include structured logs with correlation IDs, at least one observability platform such as Datadog, CloudWatch, or GCP Logs, and an alert delivery channel such as Slack or PagerDuty.<\/p>\n<p>Edge cases to plan for fall into three categories, all stemming from gaps in telemetry coverage: async message queue failures where trace propagation breaks across service boundaries, incidents caused by infrastructure changes not captured in deployment event streams, and alert storms where <a href=\"https:\/\/hud.io\/blog\/automated-root-cause-analysis-lowers-mttr\" target=\"_blank\" rel=\"noindex nofollow\">over-correlation groups unrelated problems into a single incident, creating confusion for responders<\/a>. Each scenario limits the signal available to the automated tool and makes accurate conclusions harder.<\/p>\n<h2>Struct\u2019s Ecosystem Fit for Seed\u2011to\u2011Series C Teams<\/h2>\n<p>Struct is purpose-built for Seed-to-Series C engineering teams that need immediate time-to-value without enterprise procurement cycles. It connects to Slack, GitHub, Datadog, Sentry, AWS CloudWatch, GCP Logs, Azure, Grafana, Prometheus or Loki, Sumo Logic, Better Stack, PagerDuty, Linear, and Jira. <a href=\"https:\/\/www.producthunt.com\/products\/struct-2\" target=\"_blank\">Struct is fully SOC 2 and HIPAA compliant<\/a>, which covers the compliance requirements of most fintech, healthtech, and enterprise SaaS companies at this stage.<\/p>\n<p>Custom runbooks can be pasted directly into Struct\u2019s configuration, which encodes the tribal knowledge of senior engineers into every automated investigation. Composable widgets ensure that specific visual data, such as the charts a senior SRE always pulls first, appear automatically for defined alert types.<\/p>\n<p><a href=\"https:\/\/cal.com\/deepanm\/struct-demo\" target=\"_blank\"><strong>Encode your team\u2019s investigation logic in a 30-minute demo<\/strong><\/a> and see how Struct turns tribal knowledge into automated runbooks.<\/p>\n<h2>Evaluating Automated RCA and Watchdog Alternatives<\/h2>\n<p>Teams comparing Watchdog RCA alternatives or proactive RCA tools can apply a consistent set of evaluation criteria.<\/p>\n<ul>\n<li><strong>Time-to-first-investigation:<\/strong> The tool should run an automated investigation on day one without a training period.<\/li>\n<li><strong>Cross-stack coverage:<\/strong> The platform should query all observability sources, not just a single vendor.<\/li>\n<li><strong>Interface fit:<\/strong> Output should arrive where engineers already work, such as Slack, instead of requiring a separate UI.<\/li>\n<li><strong>Runbook customization:<\/strong> The system should allow encoding of specific operational procedures.<\/li>\n<li><strong>Compliance:<\/strong> The vendor should meet data security requirements such as SOC 2 and HIPAA.<\/li>\n<li><strong>Onboarding safety:<\/strong> The tool should provide enough context for junior engineers to handle on-call independently.<\/li>\n<\/ul>\n<p>Time-to-value for AI investigation platforms is maximized when they query existing observability data directly rather than requiring a separate data pipeline. Teams should avoid tools that demand weeks of ML model training before delivering useful output.<\/p>\n<h2>Frequently Asked Questions<\/h2>\n<p><strong>What is the difference between Datadog Watchdog RCA and a proactive automated root cause analysis tool like Struct?<\/strong><\/p>\n<p>Datadog Watchdog RCA correlates anomalies detected within the Datadog platform and surfaces findings in the Datadog UI. It remains bounded by what Datadog ingests and requires engineers to navigate to a separate interface to review results. A proactive tool like Struct automatically triggers a cross-stack investigation the moment an alert fires, querying Datadog, CloudWatch, GCP Logs, Sentry, GitHub, and other sources at once, then delivering the complete root cause summary inside the Slack thread where the incident is already being managed. The engineer receives a finished investigation instead of a starting point.<\/p>\n<p><strong>How quickly can Struct be set up, and does it require dedicated engineering time?<\/strong><\/p>\n<p>Struct connects in under 10 minutes. The setup process involves authenticating three categories of integration: the alert source such as Slack or PagerDuty, the code repository such as GitHub, and the observability context such as Datadog or CloudWatch. No data pipeline configuration, ML model training period, or dedicated engineering sprint is required. The first automated investigation runs immediately after connection.<\/p>\n<p><strong>Is Struct appropriate for teams where junior engineers handle on-call rotations?<\/strong><\/p>\n<p>Struct is designed to support teams where junior engineers share on-call responsibilities. When a junior engineer receives an alert at 3 AM, Struct has already completed the initial investigation, correlated logs, mapped the timeline, assessed blast radius, and suggested a fix before the engineer opens their laptop. This removes the dependency on tribal knowledge held by senior engineers and provides a reliable, heavily contextualized starting point for every alert. Teams can safely expand their on-call rotation without requiring every responder to have deep systemic context.<\/p>\n<p><strong>What observability and alerting tools does Struct integrate with?<\/strong><\/p>\n<p>Struct integrates with Datadog, Sentry, AWS CloudWatch, Google Cloud Platform Logs, Azure Logs and Traces, Grafana, Prometheus, Loki, Sumo Logic, and Better Stack for observability context. For alert triggers, it connects to Slack, PagerDuty, Linear, Jira, and Asana. Code context comes from GitHub. This breadth of integration allows Struct to correlate signals across the full engineering stack instead of reasoning over a single platform\u2019s data.<\/p>\n<p><strong>What happens after Struct identifies the root cause?<\/strong><\/p>\n<p>After the root cause is confirmed, Struct supports a direct handoff to remediation. Engineers can pass the investigation context to a local CLI, an AI coding agent, or trigger PR creation directly from the Struct interface. This keeps the entire workflow, from alert and investigation to root cause and fix, connected without requiring engineers to manually re-explain context to a separate tool.<\/p>\n<h2>Summary: When Automated RCA Becomes Operationally Necessary<\/h2>\n<p>Manual triage functions as a structural bottleneck rather than an engineering skill gap. When senior engineers spend their weeks reacting to alerts instead of shipping product, or when SLA windows are too tight for 45-minute investigations, automated root cause analysis tools shift from a nice-to-have to an operational requirement.<\/p>\n<p>Datadog Watchdog RCA offers a useful starting point for teams whose entire stack lives inside Datadog. Teams with multi-platform observability, Slack-based incident response, and a need for cross-stack code correlation benefit more from a dedicated proactive RCA tool that closes the gaps Watchdog leaves open.<\/p>\n<p>As the comparison above shows, Struct\u2019s sub-five-minute investigation time translates to an 80% reduction in triage overhead for teams at scale, which is the difference between waking an engineer at 3 AM to hunt logs and handing them a finished root cause summary. Struct pairs this speed with a 10-minute setup and an 85\u201390% helpful investigation rate, and it is built specifically for the engineering team size and velocity requirements of Seed-to-Series C companies.<\/p>\n<p><a href=\"https:\/\/cal.com\/deepanm\/struct-demo\" target=\"_blank\"><strong>Stop burning your best engineers on 3 AM log-hunting<\/strong><\/a>. Connect your integrations in 10 minutes and let Struct handle the next investigation before you finish reading the alert.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>Learn how Watchdog RCA tools automate incident triage. Struct delivers complete root cause hypotheses in under 5 minutes \u2014 no log-hunting needed.<\/p>\n","protected":false},"author":73,"featured_media":743,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"inline_featured_image":false,"footnotes":""},"categories":[1],"tags":[],"class_list":["post-744","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-uncategorized"],"_links":{"self":[{"href":"https:\/\/struct.ai\/articles\/wp-json\/wp\/v2\/posts\/744","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/struct.ai\/articles\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/struct.ai\/articles\/wp-json\/wp\/v2\/types\/post"}],"replies":[{"embeddable":true,"href":"https:\/\/struct.ai\/articles\/wp-json\/wp\/v2\/comments?post=744"}],"version-history":[{"count":0,"href":"https:\/\/struct.ai\/articles\/wp-json\/wp\/v2\/posts\/744\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/struct.ai\/articles\/wp-json\/wp\/v2\/media\/743"}],"wp:attachment":[{"href":"https:\/\/struct.ai\/articles\/wp-json\/wp\/v2\/media?parent=744"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/struct.ai\/articles\/wp-json\/wp\/v2\/categories?post=744"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/struct.ai\/articles\/wp-json\/wp\/v2\/tags?post=744"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}