nightingale

What's Changed

• feat: native Jira, Discord, JSM Alert, Slack and Mattermost notification channels (#3408), and notification channel configs can reference user variables (#3349)
• feat: dashboard panels can mix datasources — each query target carries its own datasource, backed by a new v2 multi-datasource batch query API with KQL support (#3292)
• feat: anonymous, time-limited dashboard sharing links, plus a read-only fullscreen view with URL switches to hide the header and variables (#3345)
• feat: Grafana dashboard import rewritten — a conversion report shows what was migrated and what is unsupported before saving
• feat: dashboard query variables support label/value options and Doris SQL time macros; embedded Grafana (iframe) dashboards sync the time range and refresh interval
• feat: user accounts can be disabled and re-enabled without losing their roles, teams and permissions (#3395)
• feat: OAuth2 login can assign default teams to new users (#3406), and /login?sso=xxx redirects straight to the given SSO provider
• feat: built-in content supports arbitrary languages; ja / ru / pt-BR / es / id / ko / fr added on the backend, and pt-BR / es / id / ko / fr added to the console (#3352)
• feat: custom alert severity names, a fullscreen live view for active alert events, and filtering historical alert events by hash (#3392)
• feat: an aggregated Nightingale AI page with renamable and shareable conversations (#3325); the AI assistant can dispatch script tasks through ibex (#3378)
• feat: integration center onboarding guide (collect template → dashboard → alert rule); new x509_cert integration, rabbitmq_exporter collect template and SNMP device health alert rules (#3394, #3348)
• feat: 403 responses render in place instead of redirecting to /403, so the URL survives and going back works
• feat: passwordless demo login entry for public demo sites, off by default (#3426) — see Notes
• perf: AI chat renders markdown incrementally, and A2A task store write amplification is cut ~30x (#3413)
• security: close an SSRF guard bypass via 6to4 / NAT64 / site-local IPv6 addresses (#3364), redact credentials embedded in datasource URLs (#3424), validate TDengine database / table names (#3423), tighten the dashboard share token scope (#3368), and upgrade vulnerable dependencies (#3419, #3422)
• fix: config key initialization now converges across instances, fixing a 401 / login loop in multi-instance deployments (#3385)
• fix: an unschedulable rule cron pattern no longer crashes n9e / n9e-edge / n9e-alert on every restart, and cron patterns are validated on save (#3407)
• fix: alert mutes that are not active yet are now synced into the engine cache (#3386), and empty mute tags are normalized (#3425)
• fix: populate the recovery time before dispatching alert events (#3382)
• fix: callback channels can notify without a message template (#3350)
• fix: disabled datasources can still be queried, and are kept out of write targets
• fix: an oversized VictoriaLogs line no longer fails the whole query (#3344)
• fix: PostgreSQL datasources use the configured database when initializing the client (#3367)

v9 highlights

• Introducing Nightingale AI: a built-in AI assistant, a Skills marketplace, LLM configuration management, and page-aware prompt recommendations
• A built-in time series database that works out of the box, so small-scale setups no longer need a separate Prometheus / VictoriaMetrics
• Redesigned navigation menu, global layout and table UI
• Fully rebuilt alert rule / mute rule / subscribe rule / notification rule forms
• Reworked notification channel management
• A new event pipeline execution tracking page that shows the processing in detail
• A new generation of log explorer with unified support for Elasticsearch, Loki and VictoriaLogs

Upgrade steps

Replace the binary and the integrations directory, then restart. To pick up the new metric descriptions in the added languages, also replace `

What's Changed

• feat: list pages keep the current page in the URL, so a refresh or a trip into a detail page brings you back to where you were; the page resets to 1 only on save or when you switch business group
• feat: the user list gains an account source column, and the last-active-time column that was missing from the column config — neither rendered nor selectable — is fixed
• perf: the logs viewer renders rows through a virtual list, so large log volumes no longer stutter
• fix: concurrent webhook / callback queue creation no longer blocks on itself and drops events
• fix: a malformed HH:MM effective time span in an alert rule is now rejected outright instead of being saved silently as an unusable config
• fix: integration component icons no longer break on Linux when an upgraded database has a directory case or spelling mismatch, and the icon route now rejects path traversal
• fix: woff2 fonts and n9e-collect-templates/*.toml collect templates were treated as frontend routes and served index.html, which broke their download
• fix: the board_busigroup table never got its composite primary key, because the tag was written outside the gorm tag
• fix: correct the normalization of LLM provider base URLs that do not end in /v1
• fix: an Elasticsearch alert rule no longer keeps a stale date_field when the selected index has no date fields
• fix: the Doris log explorer resolves flattened pseudo field names back to the real columns when filtering
• fix: event pipeline header parameters are now submitted as an object
• fix: MCP APIs go through the N9E_PATHNAME prefix in sub-path deployments
• fix: the datasource dropdown search now also matches the datasource type name
• fix: raw_message is surfaced when an API call fails

v9 highlights

• Introducing Nightingale AI: a built-in AI assistant, a Skills marketplace, LLM configuration management, and page-aware prompt recommendations
• A built-in time series database that works out of the box, so small-scale setups no longer need a separate Prometheus / VictoriaMetrics
• Redesigned navigation menu, global layout and table UI
• Fully rebuilt alert rule / mute rule / subscribe rule / notification rule forms
• Reworked notification channel management
• A new event pipeline execution tracking page that shows the processing in detail
• A new generation of log explorer with unified support for Elasticsearch, Loki and VictoriaLogs

Upgrade steps

Replace the binary and the integrations directory, then restart.

Schema changes (skip this if your database account can create tables)

Nightingale creates the tables for you. If the database account Nightingale connects with cannot create tables, you need to create them by hand — the DDL is at https://github.com/ccfos/nightingale/blob/main/docker/migratesql/migrate.sql

Notes

• The new AI assistant relies on Redis Streams, so Redis 5.0 or later is required (Redis 7.0+ recommended for Cluster users, to enable sharded Pub/Sub)
• Replacing only the binary does not enable the built-in TSDB: the [EmbeddedTSDB] section exists only in the new etc/config.toml, so carrying over an old config file leaves the feature off — copy the section across yourself. It is handled by the center process only (n9e-edge / n9e-alert / n9e-pushgw ignore it and log a warning), and the data lives on that center instance's local disk, so it suits single-instance deployments only; with multiple center replicas, use an external TSDB instead
• telegraf users need to append ignore_host=false to the metrics write address, otherwise the hosts running telegraf will no longer register into the host list
• If you run n9e-edge or n9e-pushgw, upgrade them together with n9e

What's Changed

• feat: built-in time series database — metrics pushed by Categraf are stored locally, so no external Prometheus / VictoriaMetrics is required to get started; a Prometheus datasource named embedded-tsdb is auto-registered on first start. Existing setups are respected: if any enabled Prometheus datasource is already present, the datasource is not auto-created — samples are still written locally, and you can add a Prometheus datasource pointing at the built-in TSDB to query them (#3283)
• feat: import datasources from Grafana in one click, with a per-datasource result summary (#3295)
• feat: datasource page revamp — after saving a datasource, Nightingale probes it with sentinel metrics to report which built-in dashboard / alert rule templates already have data, and guides you to your first query (#3301)
• feat: per-cycle alert evaluation records — each rule evaluation persists what it queried and how it judged to the alert engine's local disk, inspectable from a drawer on the alert rule page (#3305)
• feat: notification channel configs can be tested, and message templates previewed, before they are saved (#3302)
• feat: Categraf collect-config wizard — a new /agents/categraf/collect.sh endpoint serves the applier script, with plugin templates and metric verification in the console (#3286)
• feat: event pipeline supports mock events in try-run, time-range filtering on execution records, batch enable/disable, and a new change API (#3281, #3285)
• feat: built-in alert rule templates in integrations greatly expanded and revised, with en_US i18n (#3282)
• feat: MySQL datasources can append extra params to the DSN (#3299)
• feat: Doris datasources support doris.internal_addr for the edge process to reach FE MySQL (#3298)
• feat: alert rule list supports filtering by trigger type
• feat: active alert event list shows the value at trigger time, adds three-state tags, and remembers filter panel, auto-refresh interval and time range across visits
• feat: log explorer adds a query doc button for Loki and VictoriaLogs, lets you append a field filter straight from the logs viewer, and supports BCE BLS drilldown
• feat: dashboard TableNG panel supports a row detail drawer
• feat: host page onboarding — one-click host monitoring pack, test alert action, and next-steps cards
• feat: datasource proxy and log query access are now written to the audit log with the user and the payload
• feat: edge retries event persistence to center, with a tightened per-request timeout
• feat: new API to check whether any notification was ever delivered successfully (#3300)
• perf: speed up notification record queries and the daily cleanup, backed by two new notification_record indexes created online at startup (#3311)
• fix: deep copy events before pushing them to the queue, so a recovery event no longer overwrites a not-yet-consumed trigger event in place
• fix: a blank DingTalk app config no longer breaks group robot notifications (#3309)
• fix: pin explicit column types on the ai_* models to stop AutoMigrate schema drift, which had been failing the migration of those tables
• fix: support max_completion_tokens for newer OpenAI models (#3306)
• fix: guard invalid label names in alert rule event_relabel_config (#3287, #3296)
• fix: accept YAML content for collect builtin payloads (#3307)
• fix: allow force-saving a datasource that fails its health check (#3304)
• fix: probe the global default datasource against the full datasource list before per-engine cluster filtering, so a default datasource owned by another cluster is no longer cleared
• fix: correct alert rule units
• fix: repair builtin skill YAML frontmatter and expose code search tools in chat
• fix: give the async table migration its own gorm session, so one failed statement no longer poisons the rest

V9 new feature

• feat: Introduce Nightingale AI — built-in AI assistant, Skills marketplace, LLM config management, with page-specifi

What's Changed

• feat: alert rules support test fire — trigger a rule on demand against real queried series and get a stage-by-stage report (query check → threshold → event → notify pipeline → notification), to verify the whole alerting chain before the condition is ever met (#3274)
• feat: log query supports VictoriaLogs (#3239)
• feat: alert mute rules support notify-only mute — the event is still generated and recorded (and tagged as muted in the event list and detail) while notifications are suppressed (#3250)
• feat: the host page supports one-click Categraf installation — collector packages ship with the server, and a ready-to-run install script is generated per platform (#3272)
• feat: the built-in /mcp endpoint now serves a fine-grained toolset (alerts, targets, datasource, mutes, notify rules, alert subscribes, event pipelines, users, metrics, logs, dashboards, roles) — read-only by default, with write tools and the toolset whitelist both requiring an explicit opt-in (#3245)
• feat: AI Skills support authorization scope — public or team-private, with editing restricted to authorized teams (#3247, #3268)
• feat: alert rule / mute / subscription / notification rule forms redesigned — section cards, right-side navigation with a live condition summary, auto-naming, quick duration presets, and scenario-based empty states
• feat: message templates and the template center support i18n, so a single template renders in the recipient's language (#3235, #3238, #3261)
• feat: alert history events can be retained for a configurable number of days (CleanAlertHisEventDay, cleaned in nightly batches), with a new composite index and cursor pagination (#3264)
• feat: the AI summary node in event pipelines can reuse a centralized LLM config instead of carrying its own (#3242)
• feat: full i18n coverage on the console
• feat: list tables support inline row actions, sortable creator and time columns, and a pagination quick jumper
• fix: aggregated alert event operations no longer fail with HTTP 414 — event IDs are now sent in the POST body as a JSON array
• fix: batch-deleting history events no longer removes events that still have an active alert (#3263)
• fix: quote the external config column so DDL works on MySQL 9.7 (#3276)
• fix: sanitize user input and detail URLs before HTML rendering (XSS hardening)

New Contributors

• feat: Introduce Nightingale AI — built-in AI assistant, Skills marketplace, LLM config management, with page-specific prompt recommendations
• feat: Redesigned navigation menu, global layout, and table UI
• feat: Refactored notification channel management
• feat: New event pipeline execution tracking page with detailed execution view

Upgrade note ⚠️

The new AI assistant relies on Redis Streams, so Redis 5.0 or later is required (Redis 7.0+ recommended for Cluster users to enable sharded Pub/Sub).

Full Changelog: https://github.com/ccfos/nightingale/compare/v9.0.0-beta.4...v9.0.0