Install

Ubuntu 22.04+, Debian 12+, RHEL/Rocky 9 are supported.

AMD64

wget https://github.com/ongridio/ongrid/releases/download/v0.13.4/ongrid-v0.13.4-linux-amd64.tar.xz
tar -xf ongrid-v0.13.4-linux-amd64.tar.xz && cd ongrid-v0.13.4-linux-amd64 && sudo ./install.sh

ARM64

wget https://github.com/ongridio/ongrid/releases/download/v0.13.4/ongrid-v0.13.4-linux-arm64.tar.xz
tar -xf ongrid-v0.13.4-linux-arm64.tar.xz && cd ongrid-v0.13.4-linux-arm64 && sudo ./install.sh

Both installers cache checksum-verified AMD64 and ARM64 Edge assets from CNB; fleet upgrades select the device architecture automatically.

What's Changed

• Fix Agent reliability and packet-capture artifact handling by @singchia in #332

Full Changelog: v0.13.3...v0.13.4

Install

Ubuntu 22.04+, Debian 12+, RHEL/Rocky 9 are supported.

AMD64

wget https://github.com/ongridio/ongrid/releases/download/v0.13.3/ongrid-v0.13.3-linux-amd64.tar.xz
tar -xf ongrid-v0.13.3-linux-amd64.tar.xz && cd ongrid-v0.13.3-linux-amd64 && sudo ./install.sh

ARM64

wget https://github.com/ongridio/ongrid/releases/download/v0.13.3/ongrid-v0.13.3-linux-arm64.tar.xz
tar -xf ongrid-v0.13.3-linux-arm64.tar.xz && cd ongrid-v0.13.3-linux-arm64 && sudo ./install.sh

Both installers cache checksum-verified AMD64 and ARM64 Edge assets from CNB; fleet upgrades select the device architecture automatically.

What's Changed

• docs(readme): add Trendshift badge by @singchia in #326
• feat: configure Agent output language and relax alert notes by @singchia in #327
• chore(release): prepare v0.13.3 by @singchia in #328

Full Changelog: v0.13.2...v0.13.3

Install

Ubuntu 22.04+, Debian 12+, RHEL/Rocky 9 are supported.

AMD64

wget https://github.com/ongridio/ongrid/releases/download/v0.13.2/ongrid-v0.13.2-linux-amd64.tar.xz
tar -xf ongrid-v0.13.2-linux-amd64.tar.xz && cd ongrid-v0.13.2-linux-amd64 && sudo ./install.sh

ARM64

wget https://github.com/ongridio/ongrid/releases/download/v0.13.2/ongrid-v0.13.2-linux-arm64.tar.xz
tar -xf ongrid-v0.13.2-linux-arm64.tar.xz && cd ongrid-v0.13.2-linux-arm64 && sudo ./install.sh

Both installers cache checksum-verified AMD64 and ARM64 Edge assets from CNB; fleet upgrades select the device architecture automatically.

What's Changed

• fix(grafana): sync runtime Loki datasource settings by @singchia in #323
• chore(release): prepare v0.13.2 by @singchia in #324

Full Changelog: v0.13.1...v0.13.2

Install

Ubuntu 22.04+, Debian 12+, RHEL/Rocky 9 are supported.

AMD64

wget https://github.com/ongridio/ongrid/releases/download/v0.13.1/ongrid-v0.13.1-linux-amd64.tar.xz
tar -xf ongrid-v0.13.1-linux-amd64.tar.xz && cd ongrid-v0.13.1-linux-amd64 && sudo ./install.sh

ARM64

wget https://github.com/ongridio/ongrid/releases/download/v0.13.1/ongrid-v0.13.1-linux-arm64.tar.xz
tar -xf ongrid-v0.13.1-linux-arm64.tar.xz && cd ongrid-v0.13.1-linux-arm64 && sudo ./install.sh

Both installers cache checksum-verified AMD64 and ARM64 Edge assets from CNB; fleet upgrades select the device architecture automatically.

What's Changed

• fix(edge): remove shell TCP self-check by @singchia in #320
• chore(release): prepare v0.13.1 by @singchia in #321

Full Changelog: v0.13.0...v0.13.1

Install

Ubuntu 22.04+, Debian 12+, RHEL/Rocky 9 are supported.

AMD64

wget https://github.com/ongridio/ongrid/releases/download/v0.13.0/ongrid-v0.13.0-linux-amd64.tar.xz
tar -xf ongrid-v0.13.0-linux-amd64.tar.xz && cd ongrid-v0.13.0-linux-amd64 && sudo ./install.sh

ARM64

wget https://github.com/ongridio/ongrid/releases/download/v0.13.0/ongrid-v0.13.0-linux-arm64.tar.xz
tar -xf ongrid-v0.13.0-linux-arm64.tar.xz && cd ongrid-v0.13.0-linux-arm64 && sudo ./install.sh

Both installers cache checksum-verified AMD64 and ARM64 Edge assets from CNB; fleet upgrades select the device architecture automatically.

What's Changed

• docs(readme): update release links to v0.12.0 by @singchia in #294
• fix(workflow): clarify notification delivery by @singchia in #275
• feat(knowledge): 文档切分与召回率评测 by @zhagtao in #272
• Revert "feat(knowledge): 文档切分与召回率评测" by @singchia in #304
• feat(knowledge): convert DOCX and PDF to Markdown for chunking by @zhagtao in #305
• fix(installer): make Edge bundle readable by non-root containers on hardened hosts by @ixchitl in #297
• fix(skill): tail_file test uses OS-specific absolute path by @anrenlx2025 in #231
• feat(edge/windows): add platform scaffolding (DPAPI, install, edgedirs) by @anrenlx2025 in #228
• Update README_ZH.md by @singchia in #315
• feat(aiops): redesign AgentLoop and add edge operator tools by @singchia in #316
• fix(deploy): run pcap parser as a private service by @singchia in #317
• fix(build): remove redundant runtime apt update by @singchia in #318
• chore(release): prepare v0.13.0 by @singchia in #319

New Contributors

• @zhagtao made their first contribution in #272
• @ixchitl made their first contribution in #297
• @anrenlx2025 made their first contribution in #231

Full Changelog: v0.12.0...v0.13.0

Ongrid v0.12.0

This release expands infrastructure management from hosts to Kubernetes and network devices, and strengthens context correlation across assistants, workflows, and observability.

Highlights

• Network discovery and SNMP polling: Edge discovers ARP and LLDP neighbours by default. Verify a candidate with SNMP v2c/v3 to add it as a managed network device, collect interface data, and place host-to-network links in topology.
• Kubernetes operations: Approved Kubernetes write operations are available in the legacy agent kernel, with the resulting actions retained in the operational record.
• Scoped investigation: Named-device LogQL and TraceQL queries are automatically constrained to the selected device, reducing unrelated evidence during RCA.
• Dynamic MCP tools: Verified MCP servers refresh the available chat tools without restarting the Manager.
• Configurable LLM timeout: Set the provider request timeout in Settings -> Assistant for slower model gateways.
• Usability fixes: Improved light-theme RCA readability, network-discovery hover state, Grafana metric links, device navigation E2E stability, and sidebar resilience when optional UI settings are unavailable.

Install

Ubuntu 22.04+, Debian 12+, RHEL/Rocky 9 are supported.

AMD64

wget https://github.com/ongridio/ongrid/releases/download/v0.12.0/ongrid-v0.12.0-linux-amd64.tar.xz
tar -xf ongrid-v0.12.0-linux-amd64.tar.xz && cd ongrid-v0.12.0-linux-amd64
sudo ./install.sh

ARM64

wget https://github.com/ongridio/ongrid/releases/download/v0.12.0/ongrid-v0.12.0-linux-arm64.tar.xz
tar -xf ongrid-v0.12.0-linux-arm64.tar.xz && cd ongrid-v0.12.0-linux-arm64
sudo ./install.sh

Upgrade

Download and extract the package for the server architecture, then run the included upgrader:

sudo ./upgrade.sh

The installer and upgrader fetch checksum-verified Edge assets for both AMD64 and ARM64, then select the matching architecture for each managed host.

Full Changelog: v0.11.1...v0.12.0

Install

Ubuntu 22.04+, Debian 12+, RHEL/Rocky 9 are supported.

AMD64

wget https://github.com/ongridio/ongrid/releases/download/v0.11.1/ongrid-v0.11.1-linux-amd64.tar.xz
tar -xf ongrid-v0.11.1-linux-amd64.tar.xz && cd ongrid-v0.11.1-linux-amd64 && sudo ./install.sh

ARM64

wget https://github.com/ongridio/ongrid/releases/download/v0.11.1/ongrid-v0.11.1-linux-arm64.tar.xz
tar -xf ongrid-v0.11.1-linux-arm64.tar.xz && cd ongrid-v0.11.1-linux-arm64 && sudo ./install.sh

Both installers cache checksum-verified AMD64 and ARM64 Edge assets from CNB; fleet upgrades select the device architecture automatically.

What's Changed

• docs(readme): update install links to v0.11.0 by @youzi-1122 in #264
• feat(installer): externalize Edge assets to CNB releases by @youzi-1122 in #265
• fix(release): harden CNB publishing and accelerate image builds by @youzi-1122 in #266
• fix(release): neutralize CNB uploader runner commands by @youzi-1122 in #267
• fix(installer): restore architecture-compatible Edge upgrades by @youzi-1122 in #270

Full Changelog: v0.11.0...v0.11.1

Ongrid v0.11.0

Ongrid v0.11.0 introduces reusable batch enrollment for non-Kubernetes devices, first-class device cluster management, and persistent rolling Edge upgrades. It also makes Manager upgrades safer on large data directories, strengthens Kubernetes inventory for large clusters, validates installer-discovered public addresses, and fixes several UI and concurrency issues.

This release turns device installation from a one-command-per-device workflow into a reusable, bounded enrollment flow and adds the fleet lifecycle around it: organize hosts into a device cluster, inspect membership and versions, run guarded rolling upgrades, leave the page while the job continues, and return later to review or retry individual results.

Highlights

Reusable batch enrollment for non-Kubernetes devices

The Devices page can now create a reusable enrollment batch and generate one installation command for multiple Linux hosts.

• A batch can be used as installation-only grouping or can automatically attach newly enrolled devices to a selected device cluster.
• The first non-Kubernetes device cluster can be created directly from the batch-enrollment flow; an existing cluster is not required beforehand.
• Each enrolled host receives its own Edge credential. Devices do not share a long-lived runtime access key.
• Enrollment batches are bounded by expiration and maximum-use limits and can be deleted when no longer needed.
• The enrollment token is shown only when the batch is created. Treat the generated installation command as a secret and distribute it only to intended hosts.
• Existing single-device enrollment and previously registered devices remain compatible.

Device cluster management

A new Infrastructure → Clusters area provides a dedicated management surface for non-Kubernetes device groups.

• View cluster member, online/offline, and active enrollment-batch counts from the cluster list.
• Open a cluster by clicking its row and manage members, installation batches, naming, and lifecycle from the detail page.
• Add existing eligible devices manually or let a reusable enrollment batch assign new devices automatically.
• Remove a member without deleting the underlying device.
• Rename a cluster, delete unused enrollment batches, and delete an empty cluster with server-side membership safeguards.
• Device pages now display consistent Host, Kubernetes, and device-cluster labels so ownership is visible without opening another page.

Persistent rolling Edge upgrades

Device clusters can now upgrade their Host Edge agents as a durable background operation instead of a browser-bound bulk request.

• Preflight separates eligible devices from devices that are already current, offline, missing an Edge association, using an unsupported OS or architecture, or blocked by a missing/invalid release artifact.
• Linux amd64 and arm64 devices use architecture-specific packages and SHA-256 verification.
• A single job supports up to 500 devices. The default rollout processes 10 devices per batch with at most 8 concurrent dispatches inside a batch.
• The next batch does not start until the current batch reaches a terminal state, limiting the number of simultaneously disrupted hosts.
• Jobs and per-device results are persisted. Closing the browser, leaving the page, or restarting the Manager does not discard progress.
• Completion requires both a new Edge registration and the target agent version; successfully sending an upgrade command alone is not considered success.
• Upgrade history includes original/target versions, architecture, batch, attempts, state, and failure reason. Failed or timed-out devices can be retried without repeating successful devices.
• The existing Edge package updater continues to use checksum validation, atomic replacement, and automatic rollback on apply failure.

Safer and faster Manager upgrades

Routine upgrades no longer recursively change ownership across accumulated MySQL, Prometheus, Loki, Tempo, and Grafana data.

• The normal path validates and repairs only bind-mount top-level directory metadata before stopping the running stack.
• Large observability data directories no longer cause long recursive chown scans and extended upgrade outages.
• If required permissions cannot be established, the upgrade stops before downtime where possible.
• sudo ./upgrade.sh --repair-permissions remains available as an explicit recovery operation for manually damaged permissions or backups that did not preserve ownership. Because it recursively scans data, it should not be used for routine upgrades.
• Existing .env, TLS certificates, and persisted data remain preserved.

Kubernetes inventory and observability reliability

Kubernetes inventory now uses resource-specific API shapes instead of decoding every workload kind through one shared status model.

• Active CronJobs no longer break an entire workload snapshot, while Job active/failed execution counts remain available.
• Deployment-owned ReplicaSets retain controller UID, revision, and creation metadata and appear as expandable rollout history; standalone ReplicaSets remain visible.
• Workloads, Pods, Nodes, and Events use server-side pagination, keeping requests bounded on clusters with 1,500+ Pods.
• Workload health calculations and stale HPA warning expiry are more accurate.
• Kubernetes telemetry preserves the Manager public URL for TLS origin matching.
• Tempo OTLP/HTTP ingestion checks and Manager-side Tempo query-readiness checks now target their correct protocols and listeners.

Installer, UI, and reliability fixes

• Public-address discovery rejects HTTP error pages and non-IPv4 response bodies instead of persisting them as ONGRID_PUBLIC_URL. Explicitly configured public URLs are validated before install and before an upgrade stops the old stack, with the validated outbound-interface address retained as fallback (#246).
• The device action menu measures available viewport space, opens above the trigger when appropriate, and remains scrollable and reachable on short screens (#254, closes #198).
• The in-memory chat-session test repository returns snapshots instead of shared pointers, eliminating a scheduling-dependent race in CI (#258).
• Core checkout, Go, Node, Helm, and Buf workflow actions were updated to maintained runtime generations (#259).

Compatibility and operational notes

• Upgrading from v0.10.2 does not require a manual database migration. New tables and fields are applied through the existing GORM AutoMigrate path.
• Existing Edge APIs, single-device enrollment, registered devices, Kubernetes management, and the previous synchronous bulk-upgrade API remain compatible.
• Cluster rolling upgrade currently targets online Linux Host Edge devices on amd64 and arm64. Ineligible devices are reported during preflight and are not sent an upgrade command.
• Manager installation remains Docker Compose only. Ubuntu 22.04+, Debian 12+, and RHEL/Rocky 9 are supported; Docker 24.0+ and Docker Compose v2 are required.
• Manager Compose recreation causes a short service interruption. Device-cluster rolling upgrades also briefly disconnect each Edge as its batch is processed.
• Kubernetes users should continue to use the cluster-specific Helm upgrade command generated by the Ongrid Kubernetes page. Helm 3.14+ is required for the generated command.

The v0.11.0 Kubernetes Chart is published at:

oci://helm.cnb.cool/ongridio/ongrid-edge

To inspect the Chart:

helm show chart oci://helm.cnb.cool/ongridio/ongrid-edge --version 0.11.0

Fresh installation

AMD64

wget https://github.com/ongridio/ongrid/releases/download/v0.11.0/ongrid-v0.11.0-linux-amd64.tar.xz
wget https://github.com/ongridio/ongrid/releases/download/v0.11.0/ongrid-v0.11.0-linux-amd64.tar.xz.sha256
sha256sum -c ongrid-v0.11.0-linux-amd64.tar.xz.sha256
tar -xf ongrid-v0.11.0-linux-amd64.tar.xz
cd ongrid-v0.11.0-linux-amd64
sudo ./install.sh

ARM64

wget https://github.com/ongridio/ongrid/releases/download/v0.11.0/ongrid-v0.11.0-linux-arm64.tar.xz
wget https://github.com/ongridio/ongrid/releases/download/v0.11.0/ongrid-v0.11.0-linux-arm64.tar.xz.sha256
sha256sum -c ongrid-v0.11.0-linux-arm64.tar.xz.sha256
tar -xf ongrid-v0.11.0-linux-arm64.tar.xz
cd ongrid-v0.11.0-linux-arm64
sudo ./install.sh

Upgrade

Manager

Download and extract the v0.11.0 package matching the target architecture, then run:

cd ongrid-v0.11.0-linux-<arch>
sudo ./upgrade.sh

The upgrade script validates and pulls all required images before stopping the current stack, preserves .env, certificates, and data, updates the Compose configuration, starts v0.11.0, and waits for /healthz.

Do not add --repair-permissions unless ownership was manually damaged, restored from an incompatible backup, or directed by troubleshooting guidance.

Kubernetes Edge

After upgrading the Manager, open the Kubernetes cluster page and run its generated one-command Helm upgrade in the target cluster. Verify the Controller, Telemetry Gateway, Metrics Scraper, kube-state-metrics, and Node Edge workloads are Ready after Helm completes.

Release artifacts

The release also includes checksum files and latest.json metadata for automated download discovery.

What's Changed

Device enrollment and fleet management

• #255 Add bounded reusable Edge enrollment batches, independent per-host credentials, batch-only grouping, and optional device-cluster assignment.
• #257 Add device cluster list/detail management, membership lifecycle, persistent rolling upgrades, progress history, retry, and Manager-restart recovery.

Kubernetes and observability

• #250 Normalize Kubernetes workload schemas, add bounded resource pagination and Deployment rollout history, correct health/warning behavior, and separate Tempo ingest from query probes.

Installation and upgrade safety

• #246 Validate discovered and configured public addresses before persisting or using them.
• #252 Remove routine recursive ownership scans and add an explicit permission-repair recovery mode. Closes #248.

UI and reliability

• #254 Keep the device action menu inside the viewport. Closes #198.
• #258 Remove the shared-pointer data race from in-memory chat-session tests.

Documentation, CI, and release

• #245 Update multilingual installation links for the preceding v0.10.2 release; no v0.11.0 runtime behavior change.
• #259 Update core GitHub Actions to maintained runtime generations.
• #260 Prepare the v0.11.0 release version.

Full Changelog: v0.10.2...v0.11.0

Ongrid v0.10.2

Ongrid v0.10.2 focuses on isolating Kubernetes telemetry workloads from the Controller, safer one-command cluster upgrades, earlier LLM configuration diagnostics, broader model compatibility, and usable device charts at very high series cardinality. It also improves alert context and native instant-messaging delivery.

This release resolves the Controller and heartbeat pressure reported in #235 and #239, restores Kimi K2/K3 chat compatibility from #219, and fixes the high-core chart layout from #238.

Highlights

Kubernetes control plane and telemetry data plane are now isolated

The Kubernetes Controller remains a single replica with one controller identity, but its default responsibilities are now limited to enrollment, heartbeat and tunnel management, Kubernetes inventory list/watch, live queries, and approved cluster actions. OTLP traffic and kube-state-metrics collection no longer share the Controller process, cgroup, or tunnel.

Key behavior changes:

• OTLP traces, logs, and metrics are handled by a dedicated Telemetry Gateway Deployment instead of the Controller.
• The existing OTLP Service DNS and ports remain unchanged (4317 for gRPC and 4318 for HTTP), so instrumented workloads do not need endpoint changes.
• Telemetry Gateway HPA is enabled by default with 2 minimum and 10 maximum replicas, 60% target CPU utilization, a 512 MiB memory target, and conservative scale-down behavior.
• kube-state-metrics is scraped by an independent Metrics Scraper that sends protobuf + Snappy Prometheus remote_write requests directly to the configured backend.
• Large KSM sample batches no longer occupy the Controller tunnel or delay heartbeat and re-registration traffic.
• Gateway and Scraper use a dedicated, least-privilege telemetry credential that is separate from the Controller identity. They do not inherit Controller action permissions.
• Telemetry endpoint and credential changes are published through a dedicated Secret; Gateway and Scraper hot-reload them without a workload rollout.
• Resource limits, health probes, queues, retries, a PodDisruptionBudget, and explicit compatibility modes are included in the Chart.

The Metrics Scraper intentionally stays at one replica in v0.10.2. Running multiple uncoordinated scrapers against the same KSM endpoint would duplicate Prometheus samples; future horizontal scaling requires leader election or explicit target/KSM sharding.

One-command Kubernetes upgrade and automatic migration

The Kubernetes page now generates one atomic Helm upgrade command for both simple image updates and topology-changing upgrades.

• Helm 3.14+ is required because the command uses --reset-then-reuse-values: new Chart defaults are loaded first, then existing user overrides are reapplied.
• An idempotent pre-upgrade Hook stops the legacy Controller KSM path before the new Metrics Scraper starts, preventing overlapping writes to the same time series.
• The stable OTLP Service is handed from the embedded Collector to the standalone Gateway without requiring workload configuration changes. A brief client reconnect can occur during the transition.
• --wait --wait-for-jobs --atomic waits for the migration and workloads; a failed upgrade automatically restores the previous Helm release.
• Once a cluster has migrated, later image-only upgrades perform state checks without patching or unnecessarily restarting the Controller.
• Explicit rollback modes remain available: telemetryGateway.mode=embedded and kubernetesMetrics.mode=controller restore the v0.10.1 data paths.

Validate LLM settings before they are saved

The LLM settings page can now validate the configuration being edited with a minimal real Chat Completions request.

• Administrators can run validation manually and the same validation is performed before saving.
• Invalid configurations are rejected before persistence; leaving the API key empty continues to use the existing disable behavior.
• Errors distinguish invalid credentials, inaccessible or unknown models, quota exhaustion, rate limiting, network and TLS failures, endpoint errors, malformed responses, and upstream failures.
• Validation has a 20-second timeout, does not retry, and does not log or persist the submitted API key.
• Localized, actionable guidance is shown in the settings page, and a prior success is invalidated when the configuration changes.

Kimi K2 and K3 compatibility

Kimi models that require fixed sampling parameters can now be used for normal conversations, not only the settings probe.

• Sampling overrides are omitted for the Kimi K2 and K3 families, including K2.5, K2.6, and K3.
• If an unknown future Kimi alias returns the provider's fixed-temperature error, Ongrid retries the request without the sampling override.
• Other providers and models retain the existing temperature behavior.

High-cardinality device metric charts

Device metric charts remain usable when a host exposes hundreds of series; the regression suite covers a 496-core device.

• The plot keeps a stable height while the complete, naturally sorted legend grows below it.
• Clicking a legend entry isolates that series; clicking another switches the selection; clicking the active entry again restores all series.
• Hover tooltips show only the series nearest to the pointer instead of an unscrollable list containing every series.
• Network metric tooltip scaling remains accurate after single-series selection.

Alert and IM delivery improvements

• Host-scoped alert notifications now display device=<hostname> (<ip>) when inventory data is available, while preserving the numeric device ID in labels and incident deduplication keys (#193).
• DingTalk long-connection delivery is enabled through the official Stream SDK, including callback normalization, credential validation, session-webhook replies, and a single final response for providers without message editing (#167).
• Slack replies use Block Kit with a plain accessibility fallback, Telegram replies use sanitized HTML within provider limits, and Feishu/Lark replies use native post messages with an oversized-payload fallback.
• The Chinese settings navigation now labels bidirectional integrations as “IM”, while notification delivery integrations remain “Channels”.

Compatibility and operational notes

• New Kubernetes installs default to telemetryGateway.mode=deployment and kubernetesMetrics.mode=scraper.
• Existing Kubernetes installations should upgrade with the command generated by the v0.10.2 Manager rather than manually replacing only the Controller image.
• Metrics Scraper is single-active in this release; kubernetesMetrics.replicas must remain 1 in scraper mode.
• The Manager installation package remains Docker Compose only. Linux device Edge installation and upgrade through systemd remain supported.
• Manager upgrades preserve /opt/ongrid/.env, TLS certificates, and data under /var/lib/ongrid, but Compose recreation causes a short service interruption.

The v0.10.2 Kubernetes Chart is published at:

oci://helm.cnb.cool/ongridio/ongrid-edge

To inspect the Chart:

helm show chart oci://helm.cnb.cool/ongridio/ongrid-edge --version 0.10.2

Fresh installation

Ubuntu 22.04+, Debian 12+, and RHEL/Rocky 9 are supported. Docker 24.0+ and Docker Compose v2 are required for the Manager package.

AMD64

wget https://github.com/ongridio/ongrid/releases/download/v0.10.2/ongrid-v0.10.2-linux-amd64.tar.xz
wget https://github.com/ongridio/ongrid/releases/download/v0.10.2/ongrid-v0.10.2-linux-amd64.tar.xz.sha256
sha256sum -c ongrid-v0.10.2-linux-amd64.tar.xz.sha256
tar -xf ongrid-v0.10.2-linux-amd64.tar.xz
cd ongrid-v0.10.2-linux-amd64
sudo ./install.sh

ARM64

wget https://github.com/ongridio/ongrid/releases/download/v0.10.2/ongrid-v0.10.2-linux-arm64.tar.xz
wget https://github.com/ongridio/ongrid/releases/download/v0.10.2/ongrid-v0.10.2-linux-arm64.tar.xz.sha256
sha256sum -c ongrid-v0.10.2-linux-arm64.tar.xz.sha256
tar -xf ongrid-v0.10.2-linux-arm64.tar.xz
cd ongrid-v0.10.2-linux-arm64
sudo ./install.sh

Upgrade

Manager

Download and extract the package matching the target architecture, then run:

cd ongrid-v0.10.2-linux-<arch>
sudo ./upgrade.sh

The upgrade script validates and pulls the required images before stopping the current stack, preserves configuration and data, recreates the Compose services, and waits for /healthz.

Kubernetes Edge

After upgrading the Manager to v0.10.2, open the Kubernetes cluster page and run its generated one-command Helm upgrade in the target cluster. Confirm that Helm 3.14+ and the OCI registry helm.cnb.cool are reachable first.

After the command completes, verify that the Controller, Telemetry Gateway, Metrics Scraper, kube-state-metrics, and Node Edge workloads are Ready. On clusters with Metrics Server, also confirm that the Telemetry Gateway HPA can read CPU and memory metrics.

Release artifacts

The release workflow successfully published:

• linux/amd64 and linux/arm64 Manager/Web/Kubernetes Edge image manifests;
• the v0.10.2 OCI Helm Chart;
• AMD64 and ARM64 server packages with SHA-256 checksum files;
• latest.json release metadata.

What's Changed

Kubernetes

• #241 Separate the telemetry data plane from the Controller, add the scalable Gateway and direct-write Metrics Scraper, and provide atomic Helm migration. Fixes #235 and #239.

AI and model configuration

• #237 Validate LLM configurations before saving and return actionable failure categories.
• #242 Support fixed-temperature Kimi K2/K3 models. Closes #219.

Metrics UI

• #243 Improve high-cardinality device metric charts and Grafana-style series interaction. Closes #238.

Alerts and instant messaging

• #201 Show human-readable device identity in alert notifications. Fixes #193.
• #200 Enable DingTalk Stream delivery. Fixes #167.
• #203 Render Slack, Telegram, and Feishu/Lark replies in native formats; continues #202.
• #223 Clarify the Chinese IM settings labels.

Documentation and release

• #222 Update multilingual installation links for the preceding v0.10.1 release; no v0.10.2 runtime behavior change.
• #244 Prepare the v0.10.2 release version.

Full Changelog: v0.10.1...v0.10.2

Ongrid v0.10.1

Ongrid v0.10.1 focuses on Kubernetes observability, reliable large-cluster metric collection, safer node access, and reproducible online delivery.

Standard Kubernetes onboarding now installs the components needed for cluster-state metrics and OTLP ingestion by default. The Manager, Web, Kubernetes Edge image, and Helm chart are published as versioned multi-architecture artifacts through CNB, while the server installation package is smaller and Compose-only.

Highlights

Kubernetes observability works out of the box

• kube-state-metrics is enabled by default and uses the Ongrid CNB mirror.
• The controller OTLP gateway is enabled by default and exposes OTLP/gRPC on 4317 and OTLP/HTTP on 4318 inside the cluster.
• Instrumented workloads exported through the gateway receive the Manager-owned cluster_id resource attribute.
• The Kubernetes metrics drilldown now opens Grafana 11 Explore with the provisioned Prometheus datasource.
• The default drilldown query shows pod counts by namespace and phase instead of only displaying scrape-target health.
• Node Edge keeps node.collectorMode: "off" by default, avoiding duplicate node metric ingestion while Kubernetes state metrics are collected centrally.

Both kube-state-metrics and the OTLP gateway can still be explicitly disabled through Helm values.

Large-cluster metric collection no longer drops the entire scrape

The Kubernetes metric path now streams Prometheus text exposition instead of loading the complete response into memory.

• Per-target sample limit increased to 250,000.
• Tunnel batches are capped at 10,000 samples or 4 MiB.
• Every batch has an independent push timeout.
• Successfully delivered batches remain available when a later batch fails.
• If a target exceeds the configured limit, Ongrid forwards the accepted bounded subset and reports partial-delivery status instead of discarding the whole scrape.
• Local scrape and push counters make truncation and partial delivery observable.

Safer Kubernetes node host access

Kubernetes Node Edge now mounts the host filesystem explicitly at /host/root with HostToContainer propagation.

• Host access no longer depends on entering PID 1's mount namespace with setns.
• SYS_ADMIN and SYS_PTRACE were removed from the node container capability set.
• After initialization, the Edge process runs as UID/GID 65532 with the reduced runtime capabilities required by the existing host tools.
• Legacy /proc/1/root layouts remain compatible.

This improves compatibility with environments where ptrace, AppArmor/SELinux, or user-namespace restrictions block the previous launcher.

Versioned Helm chart from CNB OCI

The Kubernetes chart is now published at:

oci://helm.cnb.cool/ongridio/ongrid-edge

Manager-generated install and upgrade commands pin the Chart version to the running Ongrid version. A source checkout and the old local path ./deploy/kubernetes/ongrid-edge are no longer required.

The Chart now uses Helm's .Release.Namespace as the single namespace source and no longer renders a separate namespace resource. Release packages do not contain a local ongrid-edge.tgz; target clusters must be able to reach helm.cnb.cool.

To inspect this release's Chart:

helm show chart oci://helm.cnb.cool/ongridio/ongrid-edge --version 0.10.1

For installation, create the Kubernetes cluster enrollment in Ongrid and run the generated Helm command. The generated command contains cluster-specific bootstrap credentials and should be handled as a secret.

Multi-architecture cloud images and reliable release reruns

• Manager, Web, and Kubernetes Edge images are published to docker.cnb.cool as linux/amd64 + linux/arm64 manifests.
• Compose runtime dependencies use verified CNB mirrors.
• Release reruns reuse complete immutable image and Helm Chart versions instead of attempting to overwrite them.
• Missing artifacts are published normally; incomplete manifests or indeterminate registry responses fail safely.

Server package changes

The Manager installation package is now Docker Compose only.

• Manager systemd installers, native Manager/Frontier binaries, and bundled observability binaries were removed.
• Linux device Edge installation and upgrade through systemd remain supported.
• The resulting amd64 server package is about 45% smaller than the previously audited payload.
• Runtime container images are pulled from docker.cnb.cool; they are not embedded in the tarball.

Requirements:

• Ubuntu 22.04+, Debian 12+, CentOS Stream 9, RHEL/Rocky 9
• Docker 24.0+
• Docker Compose v2 (docker compose)
• At least 2 GiB RAM and 10 GiB free disk
• Network access to docker.cnb.cool; Kubernetes onboarding also requires access to helm.cnb.cool

Fresh installation

AMD64

wget https://github.com/ongridio/ongrid/releases/download/v0.10.1/ongrid-v0.10.1-linux-amd64.tar.xz
wget https://github.com/ongridio/ongrid/releases/download/v0.10.1/ongrid-v0.10.1-linux-amd64.tar.xz.sha256
sha256sum -c ongrid-v0.10.1-linux-amd64.tar.xz.sha256
tar -xf ongrid-v0.10.1-linux-amd64.tar.xz
cd ongrid-v0.10.1-linux-amd64
sudo ./install.sh

ARM64

wget https://github.com/ongridio/ongrid/releases/download/v0.10.1/ongrid-v0.10.1-linux-arm64.tar.xz
wget https://github.com/ongridio/ongrid/releases/download/v0.10.1/ongrid-v0.10.1-linux-arm64.tar.xz.sha256
sha256sum -c ongrid-v0.10.1-linux-arm64.tar.xz.sha256
tar -xf ongrid-v0.10.1-linux-arm64.tar.xz
cd ongrid-v0.10.1-linux-arm64
sudo ./install.sh

Upgrade

Download and extract the package matching the target architecture, then run:

cd ongrid-v0.10.1-linux-<arch>
sudo ./upgrade.sh

The upgrade script validates and pulls every required image before stopping the current stack. It preserves /opt/ongrid/.env, TLS certificates, and data under /var/lib/ongrid, then recreates the Compose services and waits for /healthz.

Validation

The published ARM64 package and OCI Chart were exercised end to end after release:

• fresh Compose installation completed with all 10 services running;
• /healthz and /readyz returned HTTP 200;
• an in-place v0.10.1 -> v0.10.1 upgrade preserved configuration, data, cluster enrollment, and metrics;
• the OCI Chart installed successfully on a two-node Kubernetes v1.34.9 cluster;
• Controller, kube-state-metrics, and both Node Edge pods became Ready with zero restarts;
• both nodes appeared online in Ongrid and Kubernetes state metrics reached Prometheus.

Operational notes

• A Manager restart or Compose upgrade causes connected Edge agents to reconnect automatically.
• If a Manager is rebuilt with an empty database while previously enrolled Kubernetes nodes are reused, stale host credentials may need to be removed before the Node Edge can enroll against the new Manager.
• Self-signed TLS is suitable for initial testing. Replace the generated certificate for production use.

What's Changed

Kubernetes

• #208 Restore observability drilldowns and collection defaults.
• #209 Publish the Helm chart through CNB OCI.
• #210 Mount the host root without mount-namespace setns.
• #214 Preserve bounded metrics beyond the sample limit and add batched delivery.

Release and packaging

• #211 Publish cloud images to CNB.
• #217 Reuse complete immutable images on release reruns.
• #220 Remove Manager systemd packaging and make Helm publication idempotent.
• #215 Prepare the v0.10.1 release version.
• #206 Update multilingual installation documentation.

Full Changelog: v0.10.0...v0.10.1