Changelog
Follow up on the latest improvements and updates.
RSS
new
improved
Connected infrastructure views
Inventory
- Every resource in one place — Inventory now lists cloud resources, hosts, Kubernetes objects, services, and databases, and works without an AWS account.
- Simpler filtering — A Source filter (All, each account, Telemetry) replaces the Account filter, and a single Where column replaces Region and Account.
- Grouped type cards — Type cards are grouped into AWS, Kubernetes, Hosts and services, Databases, and Other.
- Related resources — The detail panel shows what each resource connects to, with an Open in … button to jump to its page.
Kubernetes
- New Map tab — See all namespaces at a glance, open a namespace to view its workloads and Services, and click any item for details.
- Topology tab — Nodes, Pods, Deployments, StatefulSets, and DaemonSets now have a Topology tab showing what each one connects to.
Hosts
- Topology tab — Each host now has a Topology tab showing its services, its traffic, and its EC2 instance with the instance’s relations.
- Agent update hint — When an AWS account is connected, hosts running an agent older than 2.2.0 show an update hint.
new
improved
A redesigned alert group page
The alert group page has been redesigned to show why alerts were grouped, what’s affected, and what to look at next, all with less noise.
- Focused summary row— The top row now shows Status, Instances, Last update, Routed by, and Grouped by, with Correct and Wrong buttons to rate the grouping.
- Grouping explained on hover— Hover over Grouped by to see the reason, the likely root, and the shared labels.
- Clearer RCA card— Shows an overview, the findings, and Read full analysis.
- Affected resources— The first alert to fire is listed at the top, with pod and node columns. Expand any row to see its relations.
- Related— Split into four lists: nodes with alerts, connected resources, other alerts, and shared resources.
- Cleaner layout— Tabs now show counts, and the Notified field, “Why these alerts were grouped” card, and Blast radius tab have been removed. Blast radius is also removed from the alert rule page.
- Auto-refresh— The page refreshes when alerts change and once every minute.
- Linked alert rules— Alert rules in tooltips now open in a new tab.
🔍 AI Investigator
The Investigator now takes on more of the detective work during an incident, moving you from a firing alert to a confirmed root cause in fewer steps.
Every investigation is grounded in evidence, tied to the exact resources involved, and delivered as a report your team can read and act on.
- Investigate from an alert— Start a root-cause investigation directly from a firing alert group, with the alert’s context already attached.
- Evidence for and against— Each likely cause shows the signals that support it and the signals that contradict it, so you can judge the conclusion with confidence.
- Named culprits— Every finding points to the specific services, pods, and hosts it relates to, rather than general hints.
- Readable reports— Investigation results are presented as a clear, structured write-up that can be shared from start to finish.
- Trace-backed findings— Jump from an issue to the traces behind it and see what is driving the errors.
- Faster AI assistant— The AI chat responds more smoothly, works through multiple lines of inquiry in parallel, and consistently reaches a conclusion.
RUM now supports JavaScript source maps, so frontend errors point to your original code instead of minified bundles.
- JavaScript source maps— Upload build source maps so error stacks show the original files and lines.
- Upload from your build— Uploads use a CI upload key and work with Vite, webpack, or the CLI.
KloudMate Agent
The KloudMate Agent now supports per-component resource controls on Kubernetes, one-click upgrades, eBPF log-trace correlation, and GitOps-safe instrumentation.
- Kubernetes agent resource controls— Set requests, limits, andpriorityClassNamefor each component (node agent, cluster agent, and fleet manager) so agent pods behave predictably under node pressure. Kubernetes agent installation
- One-click agent upgrade— Upgrade outdated agents without re-running install scripts, with upgrade paths for Linux, Windows, Docker, and Kubernetes. Upgrade the agent
- eBPF log-trace correlation— Trace IDs are now injected into application logs, so you can move between a log line and its trace with no code changes. eBPF Observability
- GitOps-safe APM instrumentation— The agent now detects workloads managed by Argo CD or Flux and stops instrumentation rollback loops. Five coexistence approaches are documented: eBPF mode, namespace annotation, committed annotation, agent patching, and policy engine. GitOps-managed clusters
Database Monitoring now supports MongoDB Atlas managed clusters through the Atlas Admin API, with no direct database connection required.
MongoDB Atlas M10/M20 managed clusters
— Monitored through the Atlas Admin API instead of a database connection.Cluster-level metrics
— Project, cluster, process, and disk metrics.Atlas logs
— Host, audit, and access logs.KloudMate Assistant has been reworked with a new workflow and refreshed integrations that connect KloudMate to your AI tools and services in both directions.
- Reworked Assistant workflow— The Assistant has been rebuilt with a redesigned workflow.
- Refreshed MCP Integrations— The MCP Integrations tab and the Connect picker have been updated.
- Connect AI clients to KloudMate— AI clients can now connect to the KloudMate platform over MCP.
- Bring your services into KloudMate— Connect other services to KloudMate through the same MCP integrations.
Every incident can now have a dedicated space for coordinated response, without switching between tools.
- Incident War Rooms — Every incident can now get a dedicated Slack channel or Microsoft Teams chat, opened manually, from a routing rule, or on escalation. Responders are pre-invited, and the incident card lets you acknowledge, resolve, add responders, or run an RCA without leaving the room.
A new Service & Infrastructure Map brings your entire stack into one connected view, showing how services, workloads, and infrastructure relate to each other in real time.
Trace dependencies across layers, spot where alerts are firing, and move from an alert to its surrounding context without switching screens.
- Live topology— Services, Kubernetes workloads, databases, hosts, and EC2 instances are mapped and linked together in a single view.
- Alerts on the map— Pods and nodes display their active alerts directly on the map, so problem areas stand out immediately.
- Alert to context in one click— Open an alert group directly in the map to see the services and resources around it.
⚡ Workflows & Automation
Introducing
Workflows
, a new automation module for turning alerts, events, and schedules into repeatable operational runbooks.
Build automation visually, connect it to your existing integrations, and keep people in the loop when an action needs approval.
*
Build without glue code
— Start from an alert, schedule, or webhook and chain together checks, actions, and conditions.*
Branch by rules
— Use the Route
step to send a workflow down different paths based on conditions.*
Alert-driven runbooks
— Use the alert router template and automatically identify the resources affected by an alert, so remediation targets the right resource.*
Run commands on hosts
— Execute shell commands on a selected host, with clear prompts when a host name is ambiguous.*
Human-in-the-loop
— Pause a workflow to collect input or approval from specific teammates before continuing.*
Import & export
— Move workflows between environments by exporting and importing them as drafts.*
Publish in one step
— Publish a workflow and take it live in a single step, with clear guidance when something is missing.*
Connected automation
— Enable automation triggers when connecting an app, upgrade existing connections, and recover from failed or revoked logins with contextual guidance.🗄️ Database Monitoring
A new
Database Monitoring
module brings database health, performance, and query visibility into KloudMate alongside your applications and infrastructure.
Monitor databases using eBPF, OpenTelemetry, or direct SQL-based monitoring depending on the level of visibility you need.
*
Multiple monitoring approaches
— Choose zero-configuration eBPF monitoring, OpenTelemetry-based monitoring, or deep SQL-level analytics.*
Broad database coverage
— Monitor ClickHouse, Redis, PostgreSQL, MySQL, SQL Server, MongoDB (including Atlas), Elasticsearch, and Snowflake.*
Redis topology
— Visualize Redis clusters, including cluster-mode deployments.*
Query & health insights
— Explore top queries, cache hit ratios, blocked sessions, checkpoints, and service load to identify database pressure quickly.*
Deep query visibility
— Direct monitoring provides query metrics, execution plans, wait-event analysis, and schema discovery where supported.🔥 Continuous Profiling
*
Faster flamegraphs
— Profiling views now render more efficiently and use fewer requests, making large flamegraphs more responsive.*
Java Flight Recorder
— Ingest JFR profiles for deeper Java runtime profiling in KloudMate.🔔 Alerting
*
More precise resource alerts
— Alerts now appear on a specific resource only when that resource is actually firing, reducing noise from healthy instances.Questions or feedback? We’d love to hear from you. 💬
Load More
→