--- title: "October - 2026" slug: "october-2026" tags: ["AI Updates", "Insights Updates", "Alerting Updates", "Cloud Updates", "NMS Updates", "Platform Updates", "Synthetics Updates", "Traffic Updates"] status: "new" updated: 2026-11-05T20:38:18Z published: 2026-11-05T20:38:18Z canonical: "kb.kentik.com/october-2026" --- > ## Documentation Index > Fetch the complete documentation index at: https://kb.kentik.com/llms.txt > Use this file to discover all available pages before exploring further. # October - 2026 The following updates were released to the Kentik platform during MONTH 2026. ## AI / Insights ### Features - **Streaming responses and Stop button**: AI Advisor answers appear as they are generated, and the view follows the response as it grows. Claude models also show a collapsible "Reasoning" summary above the chat input. While an answer runs, Send becomes Stop in the main chat and the copilot. - **Infoblox integration**: A new **Infoblox** tab in Settings > Kentik AI (admin access) connects AI Advisor to Infoblox, with support for multiple realms. This integration allows customers of both Infoblox UDDI and Kentik to ask AI Advisor questions pertaining to their Infoblox setup. - **New models**: Claude Sonnet 5.5 is the default for the advanced tier and is generally available. Claude Haiku 5.5 is added to the basic tier. - **AI Advisor sidebar**: Sections now match Data Explorer (Sessions, Recent, AI Settings). The sidebar moves to the right, with a **Hide/Show Sidebar** button. - **Journeys v1 retired**: The old Journeys page is removed, and /v4/core/journeys redirects. Journeys predates AI Advisor and is now fully replaced by it. - **Runbook permissions**: Chat viewers without runbook access see "You don't have permission to view runbooks" instead of runbook results. This comes with full RBAC support for runbooks. - **Faster failure for slow NMS queries**: The NMS tool stops after 3 minutes and suggests how to narrow the query. - **Runbook approval**: Runbook steps that run sensitive commands ask for approval and show a preview (human in the Loop/Elicitation prompts. ### Bug Fixes - **Cut-off answers**: Final answers are no longer cut at 4,000 characters, which affected answers with inline SVGs. - **Sub-agent errors**: Sub-agents no longer fail with a "tool_choice not supported" error on newer Claude models. - **Thinking indicator**: It now stays visible while sub-agents run, and tool-calling steps no longer stay stuck as "Processing". - **Long chats in the copilot**: Opening a long chat from history now scrolls correctly. - **Export and summary order**: Exports, summaries and load order now skip sub-agent steps correctly. - **Invalid tool arguments**: Tools now return an error for unknown dimensions or invalid IDs, instead of returning wrong or empty results. - **AI Site Search**: A search with no matching pages no longer returns a 500 error. - **AI settings permissions**: AI settings endpoints now respect user RBAC permissions. ## Alerting ### Features - **Reworked Suppression**s: New create, edit and view pages with a live Match Preview. Criteria cover severities, several policies and applications (including Synthetics), plus device, site, and interface context filters. Statuses are Planned, Ongoing, Expired and Invalid, and each suppression has version history. Silences can no longer be created. - **Flow policy grouping dimensions**: Dimension grouping for flow policies was revamped. - **Device shown first in alert table**: For NMS and event alerts with more than 3 dimensions, the device now always appears first. ### Bug Fixes - **Disabled severity levels**: Disabled levels are now respected. A trigger at a disabled level activates the next lower enabled level, and an alert whose level is disabled steps down or clears. - **Mitigation edits**: Strategy, timer, and metadata changes to mitigations are now saved when you update a policy. - **Bulk notification channels**: Bulk add now applies to all flow policy thresholds. - **Device Type filter**: Flow policies with a Device Type filter now save correctly. - **Alert charts**: Short spikes show at full height, the 1-minute peak is shown, and Data Explorer opens with a valid aggregation window. Port+Proto keys, long-running event alerts and tenant charts also work. - **Cleared filter erro**r: Filtering by Cleared no longer shows "Error Occurred" when a severity was later removed from an event policy. - **Empty dimension values**: "" values are no longer linked, which avoids lookup errors. - **NMS policy edits**: Changing a policy's dimensions now opens new alerts as expected. - **Deleted agents**: Alerts for deleted agent registrations are now retired. - **Manual clears**: Manual clears now take effect immediately. - **Rolling-window rules**: Editing a rule no longer causes a late alert or drops trigger context. - **Duplicate alerts**: Conflicting active alerts with the same key are cleared silently. - **AI Advisor**: Alert tools stay available when policy names repeat. ## Cloud ### Features - **Cloud Export Status page updates**: The status page under Settings > Cloud > Cloud API Status was updated for each provider and for All Providers. The Cloud overview widget has a new "Outdated" column. - **GCP service account impersonation**: Cloud exports for GCP can now authenticate by impersonating a service account, as an alternative to a saved key. This is a new option in the GCP authentication step of the export wizard. - **AWS Cloud WAN paths on the AWS map**: The AWS map now shows paths to the possible destinations behind a core network, including across accounts. ### Bug Fixes - **AWS secondary account data**: The AWS cloud export API no longer drops data for secondary accounts. - **AWS map account filter**: The Account Filter no longer shows 0 VPCs, subnets and gateways for accounts that have them. Counts update when a company's full refresh finishes, which can take 3–5 hours for large companies. If an account fails during a refresh, it keeps its previous counts. - **Azure map label**: "Load Ballancers" is corrected to "Load Balancers". - **Azure map metrics**: Percentage-based metrics are now averaged instead of summed. Only additive metrics such as bytes are summed. - **Cloud device counts**: Device subtype counts and displays now use device stats. Restricted users no longer see cloud counts they can't access. ## NMS ### Features - **SNMP Traps tab**: Device details has a new **Traps** tab that shows traps for that device. The infrastructure devices table has a new "traps detected" filter. - **Check Point**: Hide NAT connection counts and per-pool port usage and utilization. - **Ciena**: Active alarms from the ALARM-MIB, including power-card alarms. - **Juniper**: VPN, VPLS and MPLS L3VPN VRF metrics. - **A10**: SSL, service group, real server and global stats. - **Sangfor IAG**: New device support. It covers system, CPU, memory, storage, interfaces, sessions and signature library and license status. - **CFM remote MEP state**: State and RDI for each remote MEP. Cisco IOS-XR, ADVA and Accedian pick it up. - **Radware MIBs**: Added to the MIB database. ### Bug Fixes - **Device counts**: Associated device counts and the device page now match for SNMP and streaming telemetry filters. - **Availability widget**: The NMS home availability widget now shows correct values. - **Trap template interface names**: "Interface Name" is no longer blank. Trap templates now map ifDescr to Interface Name, ifName to Interface Name (short) and ifAlias to Description, as NMS does. - **Cumulus BGP neighbors**: Duplicate BGP neighbor reports are removed. Admin status is now reported, and peer AS is normalized. - **MikroTik components**: Unnamed PCI devices are typed as modules instead of Unknown. A missing field no longer fails the whole components report. ## Platform ### Features - **Faster login and app loading**: Permission checks run in parallel and make fewer database calls when the app loads. - **Agent health checks**: The agent health status indicators checks for common causes of agents that can't register, reach Kentik, install or upgrade. Each check names the cause and says what to do. - **Bulk collection method update**: The bulk dialog on the Devices page now has more fields. You can switch devices between ICMP-only and SNMP monitoring and pick a monitoring template. ### Bug Fixes - **Share links**: Public share links no longer expose company inventory lookups or NMS device data to people who open the link. The shared page uses only global reference data and the lookups saved when the link was created. - **Password reset**: Share-link sessions can no longer start an expired-password reset. The expiry check now uses the company's configured duration and excludes SSO users. - **Sessions after role changes**: A user's session is refreshed when their role changes, so removed privileges apply right away. SSO users no longer get an unexpected 2FA prompt. - **Report categories**: Report categories can no longer be read, changed or deleted across companies. - **Agent map popups**: Fixed a stored cross-site scripting (XSS) issue in kproxy agent map popups. ## Synthetics ### Features - **Metrics tab on test results**: Ping, DNS, HTTP, page load and throughput tests get a **Metrics** tab with collapsible time-series charts. A new filter bar (source agent, target or resolver, site, ASN, label) is shared with the **Traceroute** tab. The Agent results page also shows these charts. Status-code charts now summarize with max instead of average. Available to users with the embedded_ude permission. - **Default lookback scales with test frequency**: The default time range targets about 60 points. A 15 s test defaults to 15 m, and a 5 m test to 3 h. - **Alerts Timeline column**: The **Tests** table has a new default column next to Health Timeline. Each time bucket is colored by the worst alarm severity in it. Buckets before a test's creation, or while it is pending or paused, are grayed out. If alarm history is truncated, the tooltip says so. - **Synthetics on Universal Agents is generally available**: The "Enable Synth UA Capability" company permission is removed, so every company can run Synthetics on a Universal Agent. - **Separate IPv4 and IPv6 static IPs**: The agent's static IP setting is now split by IP family. - **Site names in chart legends**: Synthetics Site dimensions render as sites in UDE legends. ### Bug Fixes - **Dual-stack traceroutes**: Traceroutes for URL and hostname tests are no longer pinned to one resolved IP, so dual-stack agents show both IPv4 and IPv6 paths. - **Agent hardware swap**: Swapping an agent now merges its metadata instead of replacing it, and keeps the deactivated agent's history. - **Bulk delete**: Bulk test deletion now validates IDs and enforces label permissions. - **Preview tests**: The **Metrics** tab and agent drill-down are hidden on preview tests - **Terraform provider**: Synthetics test settings are sent as objects, and unmanaged settings stay null on read. - **Backend stability**: Fixes for database connection handling in the Synthetics backend, with timeouts and a better health check. Traffic ## Traffic ### Features - **Scope filtering**: Data Explorer and board queries respect the user's active access scope. - **Faster reverse DNS**: Lookups have bounded timeouts. In the test view, query time dropped from about 49 s to 6–8 s. - **Faster filters**: Long device, port and CIDR lists and OR chains run faster. In the benchmark, 250 CIDR terms went from 771 ms to 17 ms. - **Smaller interface-filter queries**: Interface filters (alias, description, speed, VRF) are limited to the devices you selected. ### Bug Fixes - **Negated device filters**: "Not" device filters give correct results. A contradictory filter returns an empty result immediately. - **Interface-class filters**: These filters are now tied to the correct device. - **Device filters**: Device scoping and pattern filters, including patterns with < and >, match the older behavior. - **App protocol filters**: Dimensions that depend on app protocol now take filters into account in Data Explorer and board panels. - **Percentages in non-flow queries**: Percentages are no longer summed when results are re-aggregated. - **Invalid IP errors**: An invalid IP or CIDR returns an HTTP 400 user error.