This week in IT Operations: Oct. 5–11, 2026

Welcome to the first edition of This week in IT Operations, our weekly roundup of the news shaping the ITOps industry. Every week, we’ll curate the stories that matter most to ITOps, AIOps, Agentic ITOps, SRE, and incident management leaders. This weekly briefing will include the latest news on analyst research, major outages, product launches, funding rounds, and industry events worth putting on your calendar.

Headlining this week are two major predictions coming out of Gartner. First, Gartner forecasts that 70% of enterprises will use agentic AI to operate their IT infrastructure by 2029, up from less than 5% in 2025, and that the I&O role will shift from doing the work to supervising the agents that do it. Second, Gartner says managing token consumption is fast becoming a business priority as agentic AI costs climb, and predicts that by 2029, 60% of organizations deploying AI will have a dedicated function that maps AI cost to value.

Put those together, and the message for IT leaders is clear: AI agents are coming to help run your operations, but you’ll need to prove their value and operational impact.

In other news, AI agents showed up on both sides of the reliability ledger. GitHub and Anthropic both suffered disruptive outages, and coverage tied GitHub’s to the load AI coding agents put on its platform. Meanwhile, ServiceNow, Palo Alto Networks, Intellias, and Rootly all launched Agentic ITOps capabilities, with most of them leading with governance and human-in-the-loop controls. Investors are also weighing in on AI in ITOps, with Rein Security raising $25 million to help enterprises monitor and control AI agents.

Here’s everything that happened in ITOps, and why it matters.

This week in IT Operations at a glance

• Gartner: 70% of enterprises will use agentic AI to operate IT infrastructure by 2029, up from less than 5% in 2025.
• Gartner: By 2029, 60% of organizations deploying AI will have a dedicated function that maps AI costs to value.
• GitHub suffered a string of outages that peaked on October 7, as coverage tied the platform’s strain to AI coding agents.
• Anthropic: A faulty spend-limit check blocked some paying customers across Claude products.
• ThousandEyes reports that global network outages fell 46% week over week.

1. Gartner predicts 70% of enterprises will use agentic AI to run IT infrastructure by 2029

Gartner’s Predicts 2026: AI Agents Will Transform IT Infrastructure and Operations forecasts that 70% of enterprises will deploy agentic AI to operate their IT infrastructure by 2029, up from less than 5% in 2025. The report also says the infrastructure and operations (I&O) role will shift from doing the work to supervising the AI agents that do it. Gartner published the report in December 2025, but it’s worth revisiting as agentic ops launches pile up across the industry.

Source: Gartner

Key takeaways

• The question is how fast enterprises will adopt Agentic ITOps, not whether they will. Going from under 5% to 70% in four years means most I&O teams will make their agentic AI decisions in the next 12 to 24 months. If Agentic ITOps isn’t in your 2027 planning cycle, it should be.
• The job is changing before the tools are. Supervising agents requires different skills than running a NOC, such as setting guardrails, reviewing agent decisions, and determining when an agent has earned greater autonomy. Teams that design the operating model early will adopt more quickly and safely.
• Adoption won’t automatically mean success. Plenty of agentic projects will stall between pilot and production. In our experience, the difference comes down to the data foundation underneath the agents, not the agents themselves.

2. Gartner: AI token costs are becoming a board-level concern as agentic AI scales

In its Top Strategic Predictions for 2027 and Beyond, published September 15, Gartner says managing token consumption is becoming a business priority as agentic AI costs rise quickly, making it harder to link spending to business value. Gartner predicts that by 2029, 60% of organizations deploying AI will establish a dedicated function to map the total cost of AI to value or profit. It also predicts that by 2028, 60% of Global 500 companies will embed AI FinOps controls at inference, moving cost governance from after-the-fact reporting to real-time oversight.

Source: Gartner Newsroom

Key takeaways:

• ROI scrutiny is coming for AIOps and AI SRE tools. As agentic AI shows up on the P&L, finance teams will ask what each agent costs per incident handled, not just what the license costs.
• Measure value in operational outcomes. The cleanest way to prove AI’s value in ITOps is to tie it to numbers the business already understands, such as MTTR, bridge call hours, and engineering hours returned. Capture your baselines before you deploy, or you’ll have nothing to compare against.
• Efficiency becomes a design requirement. Agents that reason over correlated, well-connected context tend to need fewer steps (and fewer tokens) than agents sifting through raw alert noise. Reducing noise before an agent starts reasoning helps both accuracy and cost.

3. GitHub outages cap a rough week, with AI coding agents adding strain

GitHub had incidents affecting Actions and pull requests on October 5 and 6. Then, on October 7, Git operations, pull requests, Actions, webhooks, and issues all degraded, and the problem recurred later that day. Downdetector reported more than 2,500 incidents within the first hour, and Android Authority reported that service was back by about 1:03 p.m. ET. According to GitHub’s incident summary, the two write-operation outages shared one root cause and lasted about 19 minutes combined, but more than 12,000 organizations saw failed requests and 19% of Actions workflow runs failed.

DevX ties GitHub’s string of outages in 2026 to traffic from AI coding agents. GitHub has said commit volume could reach about 14 billion this year, up from roughly 1 billion in 2025.

Sources: Android Authority, GitHub Status, DevX

Key takeaways:

• AI agents are a new class of load. A roughly 14x jump in commit volume isn’t organic growth. Capacity plans built around human behavior break down when agents generate traffic at machine speed, so revisit your assumptions as agent use grows within your enterprise.
• Third-party platforms are part of your incident surface. When GitHub goes down, CI/CD pipelines, deployments, and even the rollout of fixes stall. Mapping critical SaaS dependencies into your service topology helps responders answer “is it us or them?” in minutes, not hours.
• A same-day recurrence is a learning problem. The second outage shared the first one’s root cause. The faster your team captures what happened and what worked during an incident, the faster the next response starts from knowledge instead of from zero.

4. Anthropic spend-limit fault blocks paying customers across the Claude API, Claude.ai, and Claude Code

On October 7, a faulty spend-limit check incorrectly paused some Anthropic organizations that hadn’t actually reached their limits. Those organizations saw refused requests across the Claude API, Claude.ai, Claude Code, and other Claude products. Anthropic posted its resolved notice at 21:23 UTC, according to SessionWatcher. Earlier that day, a separate incident caused errors loading usage data on Anthropic’s platform usage page and Admin API usage report, making it harder for customers to check their actual spend.

Because spend limits apply at the organization level, one faulty check affected every Claude product for those organizations at once. That matters more and more as ITOps workflows come to depend on LLM providers.

Source: SessionWatcher

Key takeaways

• LLM providers are now tier-1 dependencies. If your incident response, service desk, or developer workflows run on an AI provider, treat that provider like any other critical service. Monitor its status, map which workflows depend on it, and include it in your runbooks.
• Control-plane failures can hit wider than outages. A billing or quota check appears before every request, so a single bad check can block multiple products at once. Plan for graceful degradation, so your teams can keep working, with humans back in the loop, if an AI service goes dark mid-incident.
• Usage visibility is part of reliability. The same-day usage-reporting incident echoes Gartner’s token-cost warning. Keep your own telemetry on AI consumption rather than relying only on a vendor’s dashboard.

5. ThousandEyes: global network outages fell 46% week over week

In Network World’s weekly internet health update on October 8, Cisco ThousandEyes counted 303 global network outage events for September 28 through October 4, down from 559 the week before. The count covers ISPs, cloud service provider networks, collaboration apps, and edge networks such as DNS and CDNs. U.S. outages fell 59%, from 406 to 168. The drop followed an unusually busy late September.

Source: Network World

Key takeaways

• Don’t over-read a single week. Weekly outage counts can swing by hundreds of events. Track the trend over months, and compare it with your own incident volume, before drawing conclusions about the health of the internet or your providers.
• Your users experience your outages as their own. ISP, cloud, and edge disruptions often first appear as application tickets. Correlating external network events with your internal alerts keeps teams from opening a war room for a problem they can’t fix.
• Make “is it external?” a standard first question. With DNS, CDN, and security-as-a-service providers in the count, a quick check for third-party causes should be a routine step every time you declare a major incident.

Come meet BigPanda at Gartner IOCS 2026

That’s all the news for this week! If you’re ready to see Agentic ITOps in action, come meet BigPadna at Gartner IT Infrastructure, Operations & Cloud Strategies Conference (IOCS) North America 2026, from December 8–10 in Las Vegas, NV. Stop by to learn how enterprises are using Agentic ITOps to improve MTTR, reduce L1 and MSP spend, prevent escalations, and build resilient IT operations.

On the calendar: upcoming ITOps events

A busy stretch of industry events is coming up. Here’s what to watch.