You're reading the Sunday, November 8, 2026 edition. Today's briefing →
Toronto
Sunday, November 8, 2026No. 147

Digital Plumber

Plumbing the information age

AI-curated intelligence for people who run networks. Daily coverage of AIOps, network automation, agentic operations, AI infrastructure, security and the vendors shaping them.

Today's 3 things that matter

Picked by the AI editor
  1. Security·Primary source

    eMudhra intermediate CA keys generated with three bits entropy

    Eight intermediate CAs operated by eMudhra had their private keys compromised with catastrophically weak entropy—as few as 3 bits instead of the required 384 bits for NIST P-384.

    Why it matters TLS traffic encrypted with eMudhra certificates is trivially decryptable; infrastructure teams must identify affected certificates and plan immediate replacements before browser distrust dates—especially for enterprise VPN and load balancers.

  2. Research·Primary source

    IETF publishes RFC 10069 for multipath QUIC connections

    RFC 10069 specifies a multipath extension for QUIC enabling simultaneous usage of multiple paths for a single connection.

    Why it matters Multipath QUIC is now a Proposed Standard, enabling load balancing and failover across heterogeneous paths without application changes or OS kernel modifications—critical for mobile and edge-cloud traffic engineering.

  3. Research·Primary source

    OTel-Arrow phase two cuts telemetry pipeline network overhead

    OTel-Arrow Phase 2 extends OTAP transport protocol with the OTel-Arrow Dataflow Engine, a Rust runtime using OTAP as primary in-pipeline representation.

    Why it matters Arrow-based OTAP pipelines achieve roughly 20x throughput improvement over OTLP in benchmarks; format conversion consumes majority of Collector CPU time, not processing logic. This cuts observability pipeline infrastructure costs and enables higher-cardinality telemetry without backend cost explosion.

Today's briefing

What happened, and why it matters

16 stories · 11 topics · Updated 11:43 AM ET

Datadog ships Scout 1.0 model for autonomous incident investigation

Datadog investor relations · Nov 4, 2026 · Vendor release

Datadog introduced Scout 1.0, its first proprietary AI model, integrated into Bits Investigation on November 4, 2026 to support faster incident response and root cause analysis. The platform combines native OpenTelemetry ingestion with pipeline-level control and analytics on historical data resolved into a live map of services, dependencies and changes.

Why it matters Native AI observability enables teams to tie frontend sessions to backend traces on synchronized timelines and identify affected users by impact rather than alert volume, accelerating incident correlation across silos.

Scout 1.0 represents Datadog's shift from general-purpose LLM agents to domain-specific SRE models trained for network, infrastructure and application troubleshooting. The model powers Bits Investigation, Datadog's AIOps tool for autonomous incident detection and response within defined guardrails. This launch aligns with Datadog being named a Forrester Wave Leader in AIOps Platforms Q4 2026 and positions the unified observability platform as an all-in-one AIOps engine. For NetOps and SRE practitioners, Scout 1.0's integration enables faster correlation across observability silos: it can surface why a network latency spike affected particular user sessions without manual timeline stitching. The platform received top scores in Forrester's Innovation and ModelOps/AI Governance criteria, highlighting vendor differentiation in production-grade autonomous operations—a key concern for enterprises deploying agentic systems in critical infrastructure.

Read the original at investors.datadoghq.com ↗

Broadcom uses topology-aware correlation to untangle post-merger networks

Broadcom Network Observability blog · Nov 6, 2026 · Primary source

Broadcom published a case study demonstrating how topology-aware root cause analysis and multi-vendor event correlation filters extraneous noise to isolate true origins of network degradation and pinpoint device and link dependencies during post-acquisition network integration.

Why it matters Topology-aware root cause analysis reduces MTTR by automatically correlating multi-vendor device events, suppressing extraneous symptom alarms, and pinpointing exact failure locations across hundreds of thousands of endpoints.

Broadcom demonstrated how modern network observability platforms apply event correlation across disparate telemetry to accelerate fault isolation in complex merged networks. The case study focused on a global telecom using symptom alarm suppression—a core NOC practice—to move from reactive troubleshooting to automated network intelligence. Key technical takeaway: multi-vendor event correlation requires topology awareness; the platform correlated device events across hundreds of thousands of endpoints by mapping dependencies and filtering noise. This addresses a critical NetOps challenge: when acquisitions consolidate networks from different vendors (Cisco, Juniper, Broadcom gear), traditional threshold-based monitoring generates alert storms. Event correlation backed by topology transforms this from alert volume reduction into root cause pinpointing. For SRE and NOC practitioners managing large infrastructure estates, this signals the maturity of topology-driven correlation as table stakes.

Read the original at networkobservability.broadcom.com ↗

CoreWeave extends AI backbone to seven new markets

StorageReview · Nov 5, 2026 · Industry news

CoreWeave brought 22.5MW of IT capacity online at Copenhagen's DEN01 data center and extended its Global AI Network backbone to seven Asia Pacific and Middle East markets, with 51 active data centers globally and 400Gbps backbone connectivity.

Why it matters Multi-region 400Gbps backbone enables distributed AI training and inference with reduced WAN latency, supporting sovereign cloud requirements and data residency compliance for geographically dispersed AI workloads.

CoreWeave announced on November 5, 2026 that it operationalized 22.5MW of IT capacity at Copenhagen's DEN01 data center, marking its first Denmark deployment. The company also extended its Global AI Network backbone to seven additional markets in Middle East and Asia Pacific regions, signaling multi-geography infrastructure strategy competing with hyperscalers. CoreWeave's backbone currently connects 51 active data centers globally across North America, Europe, APAC, and MENA, with eight new sites added in 2026. The 400Gbps backbone architecture enables multi-region AI training and inference with reduced WAN latency. CoreWeave combines Kubernetes-native orchestration with InfiniBand and RoCE networking optimized for GPU clusters. For DevOps and SRE teams, this geographic footprint supports sovereign cloud requirements, data residency compliance, and distributed training topologies where model sharding or asynchronous updates across regions become viable. The Copenhagen launch uses NVIDIA Blackwell systems, signaling CoreWeave's targeting of next-generation compute alongside network infrastructure. Neocloud operators like CoreWeave continue to differentiate on networking and low-touch provisioning for AI labs and enterprises building private AI infrastructure.

Read the original at storagereview.com ↗

Local services fail when subsea cables cut hidden dependencies

APNIC Blog · Nov 5, 2026 · Analysis

When international links fail, hidden dependencies can take local services with them. Based on INNOG 9 findings, the post highlights services that fail during subsea cable outages because they depend on infrastructure outside the local economy, and recommends strengthening local DNS infrastructure and identifying external dependencies before the next outage occurs.

Why it matters Network operators should audit which locally-delivered services fail when international connectivity drops, then reduce external dependencies—a practical checklist for building true infrastructure resilience.

During international connectivity loss (subsea cable repair, geopolitical shutdown, or routing failure), local services often fail even though the local network remains intact. Examples: DNS resolvers hosted outside the country, cloud-based application components, remote APIs, time synchronization systems, and mobile payment platforms. Services that could otherwise continue operating locally may fail because they depend on infrastructure outside the economy. Recommendations focused on reducing these external dependencies, including strengthening local DNS infrastructure and planning for extended outages.

The author illustrates this with a July 2024 Internet shutdown experience where it became impossible to distinguish which tools had failed due to network unavailability versus those that had failed because they relied on systems outside the country. During repair work on SEA-ME-WE 5 in April 2026, users experienced slower connectivity and some interruptions, while SEA-ME-WE 4 remained in service. For operators, the post recommends: identify services that depend on infrastructure outside the economy, deploy appropriate local DNS infrastructure, and test how services behave when international connectivity is unavailable.

Read the original at blog.apnic.net ↗

Taiwan sets November repair window for EAC1 cable

SubTel Forum · Nov 4, 2026 · Industry news

Taiwan Maritime Port Bureau announced that Taiwan has scheduled repair operations for faults affecting East Asia Crossing 1 (EAC1), section C, in waters off eastern Taiwan from November 10 to December 20 after an August cable fault. The notice establishes a repair work window and restricts vessel entry to the operation area but does not identify a repair vessel or repair method.

Why it matters Network operators depending on EAC1 capacity should understand the repair schedule and plan traffic rerouting via alternate cables; this represents a 40+ day window of reduced redundancy in Asia-Pacific connectivity.

Taiwan Maritime Port Bureau released a Notice to Mariners on November 4, 2026. The planned work window runs from November 10 through December 20, 2026. The announcement establishes a repair schedule, but does not confirm that service has been restored. The principal operational requirement is a prohibition on all vessels entering the operation area during the work. The announcement was issued October 30, 2026, by Global Optical Network Telecommunications Co., Ltd. It does not identify a repair vessel or describe how the cable will be repaired. The fault occurred on section C in the area off eastern Taiwan's coast. EAC1 is part of a submarine network ring across North Asia that was commissioned in 2005, using DWDM optical technology for high-capacity international connectivity.

Read the original at subtelforum.com ↗

eMudhra intermediate CA keys generated with three bits entropy

eMudhra official statement; Google, Apple, Mozilla browser root program updates · Nov 5, 2026 · Primary source

Eight intermediate CAs operated by eMudhra had their private keys compromised with catastrophically weak entropy—as few as 3 bits instead of the required 384 bits for NIST P-384. Google Chrome and Mozilla Firefox set SCTNotAfter of November 16, 2026; Apple set G1 removal for November 30, 2026.

Why it matters TLS traffic encrypted with eMudhra certificates is trivially decryptable; infrastructure teams must identify affected certificates and plan immediate replacements before browser distrust dates—especially for enterprise VPN and load balancers.

In October 2026, security researchers discovered that Indian Certificate Authority eMudhra Limited issued eight intermediate TLS certificates bearing guessable private keys—one famously just the digit 4. Instead of the required 384-bit P-384 keys, some had as few as 3 bits of entropy. OffSec researchers confirmed the eight P-384 keys were recoverable by testing fewer than 256 candidate values, enabling adversary-in-the-middle attacks on any TLS session anchored to those certificates. Google Chrome will no longer trust TLS server authentication certificates validating to emSign Root CA - G1 whose earliest Signed Certificate Timestamp (SCT) is dated after November 16, 2026 11:59:59 PM UTC in Chrome 153 and higher. Certificates with SCT on or before that date remain unaffected. Infrastructure teams running eMudhra certificates must audit their PKI, identify affected intermediates, and migrate to replacement CAs before browser distrust dates. eMudhra announced withdrawal from the SSL/TLS business. This represents a systemic PKI supply-chain failure requiring immediate inventory and remediation across all network edge appliances, VPNs, and load balancers.

Read the original at emudhra.com ↗

Palo Alto Networks nears $500M deal for Clutch Security

TechCrunch · Nov 5, 2026 · Industry news

Palo Alto Networks is negotiating to acquire Clutch Security, an AI agent security startup, for approximately $500M in cash plus milestone payments. Clutch is approaching $10M ARR, valuing the deal at roughly 50x ARR.

Why it matters Clutch brings specialized AI agent behavior monitoring and anomaly detection, addressing a critical gap in Palo Alto's product lineup as enterprises demand comprehensive AI security solutions.

Palo Alto Networks is in discussions to acquire Clutch Security, a three-year-old startup that secures AI agents' access to company systems, for $500M in cash plus milestone-based payouts. Clutch is currently approaching $10M in annual recurring revenue. If completed, this would be Palo Alto Networks' seventh acquisition this year, reflecting an aggressive push to build AI security depth. Like competitor Oasis Security (acquired by Cyera for $1B), Clutch focuses on managing AI agents' access to systems and provides monitoring of AI agent behavior, anomaly detection, and prevention of unauthorized data access or actions. The valuation reflects companies' growing focus on securing software against bad actors using AI to breach systems more easily. The deal signals market maturation in AI security—vendors now see protecting autonomous systems as a major enterprise priority requiring specialized, purpose-built capabilities rather than bolt-on solutions.

Read the original at techcrunch.com ↗

Five9 to buy Level AI for $325 million

Five9 SEC Filing · Nov 5, 2026 · Primary source

Five9 has signed a definitive agreement to acquire Level AI for $325M in all-cash, expected to close in Q4 2026. Level AI develops conversation analytics, agent-assist, and automation tools for customer experience.

Why it matters The acquisition marks a pivot toward buying AI capabilities and allows Five9 to sell AI tools to enterprises not yet using its cloud platform, widening market reach beyond traditional contact center operators.

Five9 announced on November 5, 2026 its acquisition of Level AI, which will augment Five9 capabilities with Level AI's native AI applications for customer experience that improve agent performance, automate interactions across voice and digital channels, and convert customer conversations into organizational insights. Level AI serves enterprises including Gusto, Chime, Empyrean and VistaPrint, handling more than one billion customer interactions annually through its proprietary Latitude family of AI models built specifically for customer conversations. The companies have collaborated since 2022, with Level AI applications already available through Five9's CX Marketplace. This acquisition places Five9 at the center of a wave of customer-service AI consolidation, as platform vendors race to own the AI layer around customer interactions—Salesforce acquired Cresta, while Dialpad and Ada announced a merger. For practitioners managing CX platforms, the deal signals that AI-native incumbents are being strategically absorbed into larger platforms, consolidating the competitive landscape.

Read the original at sec.gov ↗

Z Squared adds CTO Jeffery Harris to board

PR Newswire / SEC Filing · Nov 3, 2026 · Primary source

Z Squared Inc., a computing infrastructure company expanding into AI infrastructure, elected Chief Technology Officer Jeffery Harris to its Board of Directors on November 3, 2026. Harris, who previously served as CEO of Paradox Infrastructure and CTO of Paradox Data, brings deep expertise in high-density computing, power infrastructure, and immersion-cooled systems designed for AI workloads.

Why it matters Elevates AI infrastructure technologist to board-level oversight of capital deployment and site selection decisions as companies race to build data center capacity for AI compute.

Z Squared, a Nasdaq-listed computing infrastructure company pivoting toward AI data center operations, elected its Chief Technology Officer Jeffery Harris to the board effective November 3, 2026, filling a vacancy left by Michelle Burke's May resignation. Harris brings rare hands-on experience: he previously led Paradox Infrastructure as CEO from April 2022 to May 2025, then became CTO of Paradox Data (which Z Squared acquired in September 2026), where he designed high-density, immersion-cooled digital infrastructure for data-intensive computing. His background spans FPGA, GPU, and ASIC infrastructure design from 2012 onward. By adding its chief technologist directly to the board, Z Squared signals that power, cooling, and density constraints—the core challenges in AI infrastructure—will now inform board-level capital and site-selection decisions. The board remains five members with three independent directors under Nasdaq rules. Harris receives no additional compensation for the board role and was not appointed to any committee.

Read the original at sec.gov ↗

Versa says two-thirds of network tasks remain manual

Versa Networks · Nov 5, 2026 · Vendor release

Autonomous AI agents independently troubleshoot, optimize performance, and contain security threats in real time under defined policy guardrails. Nearly two-thirds of network tasks remain manual today, with agentic AI reducing MTTR from hours to minutes.

Why it matters Operators can shift from manual triage to autonomous incident response with guardrails, freeing engineers to focus on strategy while agents handle real-time network decisions.

Versa's post distinguishes agentic AI from traditional assistants: agents operate autonomously with memory, planning, and tooling capabilities, monitoring and deciding on actions independently. Unlike suggestion-based AI, these agents can dynamically adjust bandwidth allocation, predict hardware failures, and flag degraded paths before user impact. The piece notes that 84% of IT and network operations leaders surveyed expect to reach an AI-led operating model within 12 months. Key differentiator: agentic AI agents operate under policy guardrails and continuously monitor performance, executing changes in real time. This represents a shift from reactive alerting to proactive autonomous remediation—critical for enterprises managing AI-driven workloads that generate 450% more network traffic than human-driven applications. The governance model remains human-centric: operators set goals and guardrails while agents execute within defined boundaries.

Read the original at versa-networks.com ↗

Bell and Singtel lead carrier AI infrastructure spending

TelecomLead · Nov 8, 2026 · Industry news

Bell Canada, Ooredoo, China Mobile, China Telecom and Singtel are expanding investments in AI infrastructure and GPU computing, with disclosed commitments including C$1.7 billion from Bell Canada and S$130 million from Singtel subsidiary NCS for AI development. Singtel combines customer-backed AI cloud investments with measurable operational applications including AI-assisted customer support.

Why it matters Telecom operators are building sovereign AI clouds and GPU-as-a-service platforms as core strategic units, requiring practitioners to understand both network topology and AI workload placement as interdependent infrastructure problems.

Disclosed AI commitments in 2026 include C$1.7 billion from Bell Canada, $800 million from Ooredoo, RMB3.7 billion in AI network investment by China Mobile, and S$130 million from Singtel subsidiary NCS for AI development. Singtel expects group capital expenditure of approximately S$3 billion in FY2027, compared with S$2.5 billion in the previous financial year. Singtel is working with Sierra AI on its customer-service assistant, Shirley, as outlined in the company's 2026 annual report. Singtel Digital InfraCo has integrated connectivity, compute, data centre capabilities, and AI services into a unified operating model, deploying direct-to-chip liquid-cooled AI infrastructure with successive NVIDIA GPU platforms (H100, H200, GB200, GB300). By positioning AI resources at the edge of its telecommunications network, the company supports low-latency inferencing for real-time AI applications. This reflects operators treating AI infrastructure as a core strategic business unit rather than an operational optimization tool, requiring AIOps practitioners to architect for integrated network and compute orchestration.

Read the original at telecomlead.com ↗

IETF publishes RFC 10069 for multipath QUIC connections

IETF RFC Editor · Nov 9, 2026 · Primary source

RFC 10069 specifies a multipath extension for QUIC enabling simultaneous usage of multiple paths for a single connection. After handshake indicating multipath support, endpoints can use multiple paths; the extension enables clients to open new paths independently.

Why it matters Multipath QUIC is now a Proposed Standard, enabling load balancing and failover across heterogeneous paths without application changes or OS kernel modifications—critical for mobile and edge-cloud traffic engineering.

Each path maintains distinct packet number states for sending and receiving packets. Senders manage per-path congestion status independently using standard QUIC congestion control mechanisms. Multiple paths can be used simultaneously, though no guarantee of full disjoint paths exists. Path IDs are monotonically increasing and never reused after abandonment. For operations practitioners, multipath QUIC enables automatic failover and bandwidth aggregation without application-level changes. Per-path congestion control is essential for WAN optimization over heterogeneous links (lossy cellular + low-latency terrestrial). The standard handles NAT rebinding and address migration transparently, reducing complexity in mobile and multi-cloud scenarios. Implementations treat path management as transport-layer concern, leaving application code unchanged. Transport parameter negotiation enables incremental deployment. Cipher suites with nonces shorter than 12 bytes cannot be used with multipath extension.

Read the original at rfc-editor.org ↗

OTel-Arrow phase two cuts telemetry pipeline network overhead

OpenTelemetry Blog · Nov 5, 2026 · Primary source

OTel-Arrow Phase 2 extends OTAP transport protocol with the OTel-Arrow Dataflow Engine, a Rust runtime using OTAP as primary in-pipeline representation. Demonstrates significantly lower network overhead while preserving OpenTelemetry data model compatibility.

Why it matters Arrow-based OTAP pipelines achieve roughly 20x throughput improvement over OTLP in benchmarks; format conversion consumes majority of Collector CPU time, not processing logic. This cuts observability pipeline infrastructure costs and enables higher-cardinality telemetry without backend cost explosion.

OTAP reduces cost of representing telemetry by adopting columnar format (Apache Arrow) instead of row-oriented protobuf. The OTel-Arrow Dataflow Engine preserves compression advantage through processing and scales across cores while keeping overload behavior visible and contained. For production telemetry pipelines, that predictability matters as much as raw throughput. In row-oriented OTLP paths, telemetry arrives as protobuf bytes requiring deserialization-reserialization cycles; with OTAP-native paths, telemetry remains in Arrow record batches. This eliminates repeated format conversion, a hidden tax that compounds as telemetry volume increases. Phase 1 established OTAP representation and compression performance gains; Phase 2 delivers OTAP Dataflow Engine demonstrating column-oriented Arrow-based pipelines at production scale. SRE and AIOps teams can now run high-throughput observability infrastructure with lower CPU and memory footprint, enabling processing of richer telemetry before reaching backends.

Read the original at opentelemetry.io ↗

Google cuts Gemini Omni 1.2 Flash prices 30%

The Verge · Nov 5, 2026 · Industry news

Google released Gemini Omni 1.2 Flash on November 5 with 131,072-token context, 30% price reduction to $1.50/$9.00 per million tokens, and extended video generation to 15 seconds.

Why it matters For practitioners evaluating multimodal inference costs and video generation, this represents a direct price reduction on a production-available frontier model with extended capabilities.

Google released Gemini Omni 1.2 Flash on the Google API on November 5, 2026, succeeding the Omni 1.1 Flash released November 4. The model maintains a 131,072-token context window but reduces pricing by 30 percent to $1.50 per 1M input tokens and $9.00 per 1M output tokens, compared to the prior version. The key capability upgrade is extended video generation duration, now supporting 15-second video clips. This represents a direct refresh for teams evaluating multimodal video inference costs in production. The model is generally available on the Google API without waitlist restrictions, making it immediately accessible for workload evaluation and deployment.

Read the original at theverge.com ↗

Shield AI buys radar maker Echodyne for $5.3 billion

Bloomberg · Nov 5, 2026 · Industry news

Shield AI is acquiring Echodyne Corp., a radar sensor provider to government and commercial customers, for $5.3 billion while raising $2.5 billion in new funding. The funding round values Shield AI at $28.8 billion in Series H, more than doubling its March valuation of $12.7 billion.

Why it matters Capital flowing to autonomous systems with sensor fusion capabilities signals enterprise appetite for integrated defense AI amid geopolitical competition and reshuffles valuation hierarchy in physical AI.

Shield AI announced acquisition of Echodyne Corp. for $5.3 billion on November 5, funded through $2.5 billion in new capital: $1 billion Series H equity at $28.8 billion valuation, $1 billion structured equity from Blackstone, and $500 million debt from JPMorgan Chase. The valuation more than doubles Shield's $12.7 billion March 2026 mark. Deal structure includes $4.8 billion cash and equity plus $500 million conditional on Echodyne hitting milestones. Shield AI makes autonomous aircraft systems for militaries; Echodyne sells radar sensors to government and commercial customers. This represents significant M&A activity in autonomous systems despite broader AI market turbulence. The capital structure—mixing equity, structured financing, and debt—reflects investor confidence in physical AI despite regulatory headwinds against compute infrastructure expansion.

Read the original at bloomberg.com ↗

OpenAI raises $11B at $1.25T below its target

Bloomberg · Nov 6, 2026 · Industry news

OpenAI secured $11 billion at $1.25 trillion valuation, falling significantly short of earlier $30 billion/$1.4 trillion targets following a Democratic sweep of Congress and Nasdaq correction. The final round may reach approximately $20 billion, down from $30 billion initially sought.

Why it matters Democratic regulation threats and data center backlash are creating K-shaped AI capital allocation: established players with existing capacity gain advantage while debt-funded competitors face tighter financing and regulatory scrutiny.

OpenAI raised $11 billion at $1.25 trillion valuation on November 6, well below initial targets of $30 billion and $1.4 trillion valuation, following Democratic Congressional sweep and political backlash against data center expansion. AI stocks, particularly chip and data center companies, declined for second consecutive day as investors feared regulatory tightening of the AI buildout. Senator Bernie Sanders called for national data center moratorium. Democratic victories in key gubernatorial races raised prospect of stricter data center project scrutiny going forward. Goldman Sachs revised 2030 global data center capacity forecast upward to 273 GW amid uncertainty over financing and permitting timelines. The capital retrenchment signals structural shift: regulatory and political friction now fragmenting capital allocation, with incumbent tech giants holding existing capacity gaining relative advantage over debt-funded startups competing for new projects under tighter lending and permitting conditions.

Read the original at bloomberg.com ↗
Nothing in today's briefing matches that.

Listening

Podcasts and talks
  • DEV Community

    KubeCon NA 2026: Kubernetes becomes the operating platform for production AI and agents

    KubeCon + CloudNativeCon North America 2026 (Nov 9-12, Salt Lake City) debuts a new AI Inference + Agentic track, marking Kubernetes' pivot from infrastructure debate to operational discipline for production AI workloads and agentic systems at scale.

  • Packet Pushers

    HN849: Meter Acquires Packet Pushers

    Meter has acquired Packet Pushers, the 16-year-old network engineering podcast and technical blog. Guest host Scott Robohn interviews Ethan Banks and Anil Varanasi, CEO of Meter, to discuss how the acquisition will work while maintaining editorial independence.

Vendor Radar

Last 7 days · arrows compare with the 7 before

What changed this week

Last 7 days vs the 7 before

Biggest moves

Trending topics

agent · automation · observability · OpenTelemetry · MCP · Agentic AI · AIOps · RAG · SASE · SRE

Get the daily briefing

Today's 3 things that matter and every story with why it matters, in your inbox each morning. Free, and you can unsubscribe at any time. Prefer a reader? Follow the RSS feed.